Formation of palladium(0) complexes from Pd(OAc)2 and a bidentate phosphine ligand (dppp) and their reactivity in oxidative addition

Article abstract of DOI:10.1021/om0101137

A Pd⁰ complex is spontaneously generated from Pd(OAc)₂ and a bidentate phosphine such as dppp (1,3-bis(diphenylphosphino)propane). dppp is the reducing agent and is oxidized to the hemioxide dppp(O). The intramolecular reduction step is reversible. A stable Pd⁰ complex is quantitatively formed in the presence of 2 equiv of dppp, water, and a base (NEt₃). The oxidative addition of PhI gives a cationic complex, PhPd(dppp)(dppp(O))⁺, in which dppp(O) behaves as a monodentate ligand. PhPd(OAc)(dppp) is formed in the presence of added AcO^-. The oxidative addition of PhI is an intricate reaction whose kinetics has been investigated only in the presence of added AcO^-. It involves reactive dimeric or/and monomeric Pd⁰ complexes ligated by AcO^- whose relative reactivity is a function of the PhI concentration.

Full text of DOI:10.1021/om0101137

Organometallics 2001, 20, 3241-3249
3241
F or m a tion of P a lla d iu m (0) Com p lexes fr om P d (OAc)2
a n d a Bid en ta te P h osp h in e Liga n d (d p p p ) a n d Th eir
Rea ctivity in Oxid a tive Ad d ition
Christian Amatore,* Anny J utand,* and Audrey Thuilliez
De´partement de Chimie, Ecole Normale Supe´rieure, UMR CNRS 864024,
Rue Lhomond, F-75231 Paris Cedex 5, France
Received February 12, 2001
A Pd0 complex is spontaneously generated from Pd(OAc)2 and a bidentate phosphine such
as dppp (1,3-bis(diphenylphosphino)propane). dppp is the reducing agent and is oxidized to
the hemioxide dppp(O). The intramolecular reduction step is reversible. A stable Pd0 complex
is quantitatively formed in the presence of 2 equiv of dppp, water, and a base (NEt3). The
oxidative addition of PhI gives a cationic complex, PhPd(dppp)(dppp(O))+, in which dppp(O)
behaves as a monodentate ligand. PhPd(OAc)(dppp) is formed in the presence of added AcO-.
The oxidative addition of PhI is an intricate reaction whose kinetics has been investigated
only in the presence of added AcO-. It involves reactive dimeric or/and monomeric Pd0
complexes ligated by AcO- whose relative reactivity is a function of the PhI concentration.
Sch em e 2. Mech a n ism of th e Heck Rea ction
P er for m ed w ith th e Ca ta lytic P r ecu r sor : P d (OAc)2
a n d P P h 3
In tr od u ction
It has been established that a Pd0 complex is spon-
taneously formed from Pd(OAc)2 and monodentate
phosphine ligands by an irreversible intramolecular
reduction of a PdII complex by the phosphine (Scheme
1).1 Phosphine oxide is formed in the reduction process.
Sch em e 1. Mech a n ism of th e F or m a tion of P d 0
fr om P d (OAc)2 a n d P P h 3
At least 3 equiv of PPh3 are required to get a stable
anionic 16-electron complex, Pd0(PPh3)2(OAc)- (eq 2),
in which one acetate of the Pd(OAc)2 precursor remains
ligated to the Pd0 center.2 Pd0(PPh3)2(OAc)- is the
reactive species in oxidative addition of phenyl iodide
(eq 3), which yields the unexpected PhPd(OAc)(PPh3)2
complex.1,2 The latter is the reactive species in Heck
reactions, establishing the important role played by the
acetate ions delivered by the catalytic precursor (Scheme
2).2,3
a
The steps that have been fully characterized kinetically
are indicated using solid arrows, while the others are shown
with dashed arrows.3
many Pd-catalyzed reactions,4-7 among them the Heck
reactions.5-7 Chiral bisphosphine ligands may induce
asymmetric reactions.6 The spontaneous formation of
(1) (a) Amatore, C.; J utand, A.; M’Barki, M. A. Organometallics
1992, 11, 3009. (b) Amatore, C.; Carre´, E.; J utand, A.; M’Barki, M. A.
Organometallics 1995, 14, 1818.
(2) Amatore, C.; Carre´, E.; J utand, A.; M’Barki, M. A.; Meyer, G.
Organometallics 1995, 14, 5605.
(3) Amatore, C.; J utand, A. Acc. Chem. Res. 2000, 33, 314.
(4) Tsuji, J . Organic Synthesis via Palladium Compounds; Springer-
Verlag, New York, 1980.
(5) (a) de Mejeire, A.; Meyer, F. E. Angew. Chem., Int. Ed. Engl.
1994, 33, 2379. (b) Kondo, K.; Sodeoka, M.; Mori, M.; Shibasaki, M.
Synthesis 1993, 920. (c) Cabri, W.; Candiani, I. Acc. Chem. Res. 1995,
28, 2.
Addition of bidentate bisphosphine ligands to
Pd(OAc)2 also provides efficient catalytic precursors for
* Correponding authors. Fax: 33
1 44 32 33 25. E-mail:
Anny.J utand@ens.fr and amatore@ens.fr.
10.1021/om0101137 CCC: $20.00 © 2001 American Chemical Society
Publication on Web 06/21/2001

3242 Organometallics, Vol. 20, No. 15, 2001
Amatore et al.
a Pd0 complex, Pd0(Binap)2, from Pd(OAc)2 associated
with 3 equiv of Binap (2,2′-bis(diphenylphosphino)-1,1′-
binaphthyl) has been reported by Hayashi and Osawa.8
Binap has been identified as the chemical reducer,
which is oxidized to the hemioxide Binap(O). In this
case, the formation of the Pd0 complex is significantly
accelerated by addition of water.
Sch em e 3. Ten ta tive Mech a n ism for th e
F or m a tion of P d 0 fr om P d (OAc)2 a n d 1 equ iv of
d p p p
We report here mechanistic investigations on the
mechanism of formation of Pd0 complexes from Pd(OAc)2
and n equiv (n ) 1-3) of dppp (1,3-bis(diphenylphos-
phino)propane, PPh2(CH2)3PPh2)7 as a prototype of
chelating bisphosphines and on their reactivity in
oxidative addition. In contrast to PPh3, which is oxidized
to (O)PPh3 (Scheme 1), an insignificant poor ligand that
does not play any role in catalytic reactions, the situa-
tion is expected to be much more complicated in the case
of bidentate ligands. Indeed, the mono-oxidation of
PPh2(CH2)3PPh2 provides a hemioxide ligand, PPh2-
(CH2)3PPh2(O) (named dppp(O) in the following), which
still remains a monodentate phosphine ligand able to
coordinate palladium complexes.
or by reaction of PhI in an oxidative addition. Addition
of water in large excess did not induce any formation
of a Pd0 complex, which could have been detected in
cyclic voltammetry. The 31P NMR singlet of Pd(OAc)2-
(dppp) was not affected by the presence of water or PhI
added in excess (10 equiv). No new signal appeared,
which could have characterized a (phenyl)palladium(II)
complex formed by oxidative addition of PhI to a Pd0
complex. The formation of a Pd0 complex is then
thermodynamically not favored.
At higher temperature (48 °C), the solution progres-
sively turned to dark brown with formation of black
palladium, which could not be characterized either by
cyclic voltammetry or by 31P NMR. The reduction peak
of Pd(OAc)2(dppp) progressively disappeared, suggesting
that a Pd0 complex was progressively formed but was
too unstable12 because of a lack of coordinating ligand,
excluding thus any fine kinetic investigation.
Resu lts a n d Discu ssion
Evid en ce for th e F or m a tion of P a lla d iu m (0)
Com p lexes fr om P d (OAc)2 a n d d p p p in DMF .
P d (OAc)2 + 1 equ iv of d p p p . Addition of 1 equiv of
dppp to Pd(OAc)2 in DMF (containing nBu4NBF4, 0.3
M) results in the formation of a yellow-orange solution,
which exhibited a reduction peak at Ep ) -1.47 V vs
SCE.9 The complex formed in solution was relatively
stable at room temperature, as attested by its constant
reduction peak current intensity (the current intensity
is proportional to the electroactive species concentra-
tion). The complex exhibited a broad 31P NMR singlet
at 11.0 ppm (∆ν1/2 ) 30 Hz), which was assigned to
Pd(OAc)2(dppp) (eq 4).10 No oxidation peak was detected
when the cyclic voltammetry was performed directly in
oxidation, suggesting that either no Pd0 complex was
spontaneously formed from Pd(OAc)2(dppp) or that a Pd0
complex is formed but via a slow and endergonic
equilibrium (eq 5 in Scheme 3), at a very low concentra-
tion, preventing any detection by cyclic voltammetry or
31P NMR spectroscopy.
P d (OAc)2 + 2 equ iv of d p p p . The situation changed
dramatically when 2 equiv of dppp was added to
Pd(OAc)2 (2 mM) in DMF. Indeed, at room temperature,
the reduction peak current intensity of Pd(OAc)2(dppp)
at Ep ) -1.51 V vs SCE (R1 in Figure 1a) decreased
with time, while an oxidation peak O1 at Ep ) -0.41 V
was observed (Figure 1b) whose oxidation peak current
intensity increased as a function of time. This oxidation
peak disappeared in the presence of PhI; consequently
it characterizes a Pd0 complex. The formation of the Pd0
complex was also monitored by 31P NMR spectroscopy.
The magnitude of the signal of Pd(OAc)2(dppp) at 11.0
ppm decreased with time, while a new singlet was
detected at 4.20 ppm and assigned to a Pd0 complex,
which disappeared in the presence of PhI. Two other
signals were also detected at 29.6 and -17.0 ppm and
assigned to the hemioxide PPh2(CH2)3PPh2(O) by com-
parison with an authentic sample of dppp after oxida-
tion by dioxygen. The former signal characterizes the
phosphine oxide OdPPh2-, the latter the phosphine
-PPh2 of dppp(O).
The equilibrium (eq 5) could be shifted toward its
right-hand side, either by water addition11 (Scheme 3)
(6) (a) Osawa, F.; Kubo, A.; Hayashi, T. J . Am. Chem. Soc. 1991,
113, 1417. (b) Hayashi, T.; Kubo, A.; Ozawa, F. Pure Appl. Chem. 1992,
64, 421. (c) Ozawa, F.; Kubo, A.; Matsumoto, Y.; Hayashi, T.; Nishioka,
E.; Yanagi, K.; Moriguchi, K. Organometallics 1993, 12, 4188. (d)
Shimizu, I.; Matsumoto, Y.; Shoji, K.; Ono, T.; Satake, A.; Yamamoto,
A. Tetrahedron Lett. 1996, 37, 7115. (e) Shibasaki, M.; Boden, C. D.
J .; Kojima, A. Tetrahedron 1997, 53, 7371. (f) Shibasaki, M.; Vogl, E.
M. J . Organomet. Chem. 1999, 576, 1.
(7) For Heck reactions catalyzed by Pd(OAc)2 and dppp, see ref 5c
and: (a) Cabri, W.; Candiani, I.; DeBernardinis, S.; Francalanci, F.;
Penco, S.; Santi, R. J . Org. Chem. 1991, 56, 5796. (b) Cabri, W.;
Candiani, I.; Bedeshi, A.; Penco, S.; Santi, R. J . Org. Chem. 1992, 57,
1481. (c) Cabri, W.; Candiani, I.; Bedeshi, A.; Santi, R. J . Org. Chem.
1992, 57, 3358. (d) Herrmann, W. A.; Broâmer, C.; O¨ fele, K.; Beller,
M.; Fisher, H. J . Mol. Catal. A 1995, 103, 133. (e) Larhed, M.; Hallberg,
A. J . Org. Chem. 1996, 61, 9582. (f) Vallin, K. S. A.; Larhed, M.;
J ohansson, K.; Hallberg, A. J . Org. Chem. 2000, 65, 4537. (g) Qadir
M.; Mo¨chel, T.; Hii, K. K. Tetrahedron 2000, 56, 7975.
This establishes that (i) a stable Pd0 complex is
formed at room temperature from Pd(OAc)2(dppp), in
the presence of 2 equiv of dppp per Pd(OAc)2, and (ii)
dppp is the reducing agent. The 31P NMR singlet of
Pd(OAc)2(dppp) disappeared completely only after ad-
(8) Ozawa, F.; Kubo, A.; Hayashi, T. Chem. Lett. 1992, 2177.
(9) Thuilliez, A. DEA University Paris VI, J une 1998.
(10) 12.3 ppm in DMSO, see: Wehman, P.; van Donge, H. M. A.;
Hagos, A.; Kamer, P. C. J .; van Leeuwen, P. W. N. M. J . Organomet.
Chem. 1997, 535, 183.
(12) As observed by Kollar et al, a mixture of Pd(OAc)2 and 1 equiv
of dppp gave rise to an undefined Pd0 complex. Reactions were then
conducted from mixtures of Pd(OAc)2, 1 equiv of dppp, and 1 or 2 equiv
of PPh3, leading to competition between dppp and PPh3 for the
reduction of the PdII center. See: Csakai, Z.; Skoda-Fo¨ldes, R.; Kollar,
L. Inorg. Chim. Acta 1999, 286, 93.
(11) For an accelerating effect of water on the kinetics of formation
of Pd0 complexes, see ref 8.

Formation of Palladium(0) Complexes
Organometallics, Vol. 20, No. 15, 2001 3243
Sch em e 4. Mech a n ism of th e F or m a tion of P d 0 fr om P d (OAc)2 a n d 2 equ iv of d p p p in DMF
To take into account the important role of the second
dppp ligand as well as the noninfluence of water on the
rate of formation of a Pd0 complex in the presence of
only 1 equiv of dppp, a mechanism is proposed (Scheme
4), which differs from the tentative mechanism proposed
earlier in Scheme 3.
In the presence of an extra dppp, a fast reversible
ligand substitution occurs on the Pd0 complex 2, formed
in the slow intramolecular reduction step (1 f 2). The
overall reaction goes to completion by hydrolysis of the
released phosphonium salt, which produces the hemi-
oxide dppp(O) (Scheme 4).
As performed in earlier works of this group for PPh3,1
the rate of formation of the Pd0 complex from a solution
of Pd(OAc)2 (2 mM in DMF containing nBu4NBF4, 0.3
M) and dppp (2 equiv) was monitored by amperometry
at a rotating gold disk electrode polarized at -0.1 V, on
the oxidation plateau wave of the Pd0 complex. The
oxidation current intensity of the Pd0 complex (propor-
tional to its concentration) was recorded as a function
of time (Figure 2a): (i) at 48 °C to have reasonable
reaction times, (ii) in the presence of water (20 equiv)
to observe an overall irreversible reaction, and (iii) in
the presence of NEt3 (30 equiv) to avoid the decomposi-
F igu r e 1. Cyclic voltammetry performed in DMF (con-
taining nBu4NBF4, 0.3 M) at a steady gold disk electrode
(0.5 mm diameter in a and 2 mm diameter in b; scan rate
0.2 V s-1) at 20 °C. (a) Pd(OAc)2 (2 mM) and dppp (4 mM):
reduction of Pd(OAc)2(dppp). (b) Oxidation of the Pd0
complex generated from Pd(OAc)2 (2 mM) and dppp (4 mM),
120 s after mixing.
tion of the generated Pd0 complex. The plot of ln[(ilim
-
i)/ilim] versus time was linear (Figure 2b), characteristic
of a first reaction order in palladium: (ilim - i)/ilim
)
([Pd0]lim - [Pd0])/[Pd0]lim with ilim ) oxidation current
intensity at infinite time and i ) oxidation current
intensity at t. The rate constant k for the formation of
the Pd0 complex 3 (Scheme 4, where the irreversible
reaction 1 f 2 is rate determining) is then determined
from the slope of the regression line (Figure 2b).
ln(([Pd0]lim - [Pd0])/[Pd0]lim) ) -kt with k ) 2.8 × 10-3
s-1 at 48 °C. The formation of a Pd0 complex from
Pd(OAc)2 and 2 equiv of dppp is thus slightly slower
than that from PPh3 (k ) 3.6 × 10-3 s-1 at 48 °C).1b
When only 1 equiv of dppp was considered, the
reverse reaction 2 f 1 became preponderant so that no
Pd0 complex was formed at room temperature. The
reverse reaction 2 f 1 is in fact an intramolecular
oxidative addition of a phosphonium salt to a Pd0 center
(activation of the P-OAc bond in complex 2), which was
dition of water, indicative of an equilibrated reaction
shifted to completion by water (as observed by Osawa
and Hayashi in the case of Binap8), a phenomenon that
was not observed when only 1 equiv of dppp was added
to Pd(OAc)2 (vide supra). The stability of the Pd0
complex was improved in the presence of a base such
as NEt3, as observed for PPh3.2 This was interpreted
as a stabilization of the anionic complex Pd0(PPh3)2(OAc)-
vis a` vis its decomposition to the unstable Pd0(PPh3)2
by the protons (formed concomitantly with the Pd0
complex, Scheme 1) prone to react with the ligated
acetate to form acetic acid.3 By analogy, one deduces
that the Pd0 complex formed from Pd(OAc)2 and 2 equiv
of dppp is also an anionic Pd0 complex ligated by one
acetate, which is stabilized in the presence of a base
(eq 6). Moreover, in 31P NMR, it appears as a singlet,
indicative of two equivalent 31P.
1
not observed when starting from Pd(OAc)(PPh3)2 be-
cause it would then be intermolecular (Scheme 1). This
explains why the reduction step in the case of PPh3 was
irreversible (Scheme 1) and why no accelerating effect
of water was observed.1
P d (OAc)2 + 3 equ iv of d p p p . Addition of 3 equiv of
dppp to Pd(OAc)2 in DMF (containing nBu4NBF4, 0.3
M) together with 30 equiv of NEt3 and 1 equiv of H2O
results in the formation of a yellow-orange solution,
which exhibited the reduction peak of Pd(OAc)2(dppp).
When the cyclic voltammetry was performed from time
to time, the reduction current intensity decreased up
Therefore, a tentative overall reaction can then be
expressed as in eq 7, at this preliminary stage in the
study.

3244 Organometallics, Vol. 20, No. 15, 2001
Amatore et al.
Sch em e 5. Oxid a tive Ad d ition of P h I to th e P d 0
Com p lex Gen er a ted fr om P d (OAc)2 a n d 2 equ iv
d p p p in DMF
PhI to the Pd0 complex formed quantitatively in a
solution of Pd(OAc)2 with 2 equiv of dppp in the presence
of NEt3 (3 equiv) in DMF gave a (phenyl)palladium(II)
complex. The signal of the Pd0 complex at 4.20 ppm was
no longer observed as well as the signal at -17.0 ppm
characteristic of the nonoxidized PPh2 moiety of the
hemioxide dppp(O). The 31P NMR spectrum of the
(phenyl)palladium(II) complex revealed a complex struc-
ture involving three different 31P ligated to the PdII
center: 15.0 ppm (dd J PPtrans ) 359 Hz, J PPcis ) 34 Hz,
PPh2 of dppp(O) cis to Ph), 7.9 ppm (dd, J PPtrans ) 359
Hz, J PPcis ) 51 Hz, P of dppp cis to Ph), -3.1 ppm (dd,
J PPcis ) 34 and 51 Hz, P of dppp trans to Ph). An
additional singlet at 29.7 ppm characterizes the PdO
moiety of the mono-ligated dppp(O). This gives evidence
for a structure of complex 4 (Scheme 5) in which the
PdII is bis-ligated by one dppp and mono-ligated by the
hemioxide dppp(O) (formed concomitantly with the Pd0
complex) acting thus as a monodentate ligand.
F igu r e 2. Kinetics of the formation of the Pd0 complex
generated from Pd(OAc)2 (2 mM) and dppp (4 mM), under
conditions where the reaction is irreversible, i.e., in the
presence of H2O (40 mM) and NEt3 (60 mM), monitored as
a function of time at a rotating gold disk electrode (2 mm
diameter, ω ) 105 rad s-1) polarized at -0.1 V. (a)
Variation of i/ilim ) [Pd0])/[Pd0]lim with time (ilim ) oxidation
current intensity at infinite time, i ) oxidation current
intensity at t). (b) ([): experimental variation of
ln((ilim-i)/ilim) ) ln(([Pd0]lim - [Pd0])/[Pd0]lim) with time; solid
line: ln(([Pd0]lim - [Pd0])/[Pd0]lim) ) -kt with k ) 2.8 ×
10-3 s-1 at 48 °C.
to zero while the oxidation peak of a Pd0 complex at
-0.41 V was observed, its oxidation peak current
intensity increasing with time. However, this oxidation
peak did not disappear after addition of PhI (30 equiv).
The rationalization given above (Scheme 5) requires
that the 16-electron four-ligated complex 4, PhPd(dppp)-
(dppp(O))+, is a cationic complex. Its ionic character was
confirmed by conductivity measurements. The initial
conductivity due to the ionic character of the Pd0
complex (2 mM in DMF) formed in eq 7 increased from
5.4 to 39 µS after addition of PhI (10 equiv).14 The
presence of released I- was confirmed by cyclic voltam-
metry. One oxidation peak was indeed detected at +0.45
V vs SCE and readily assigned to the oxidation of free
I-, by comparison with that of an authentic sample of
nBu4NI. The structure of a related complex, p-CN-C6H4-
Pd(dppp)(dppp(O))+, very similar to complex 4, has been
observed by Hallberg and Larhed,16 based on electro-
spray mass spectroscopy during a Heck reaction from
p-CN-C6H4OTf with Pd(OAc)2/dppp as the catalytic
precursor. The value M ) 1048 implies the mono-
complexation of dppp(O), as evidenced in the present
work by different approaches.
The formation of the Pd0 complex was monitored by
31P NMR spectroscopy. The magnitude of the singlet of
Pd(OAc)2(dppp) at 11.0 ppm decreased with time, while
the singlet of the Pd0 complex at 4.2 ppm increased with
time. dppp(O) was also detected on the spectrum by its
two characteristic singlets (vide supra). As observed in
cyclic voltammetry, the Pd0 complex did not react with
PhI. This is consistent with the formation of the
coordinately saturated 18-electron complex Pd0(dppp)2
(eqs 8, 9), known to be nonreactive in oxidative additions
of aryl halides.13
The rate of formation of Pd0(dppp)2 determined at 48
°C by the amperometry technique (vide supra) was
similar to that of the Pd0 complex generated from
Pd(OAc)2 + 2 equiv of dppp. This confirms that the rate-
determining step is indeed the reaction 1 f 2 (Scheme
4), followed by the fast ligand exchange by one or two
dppp ligands according to the dppp excess.
Com p lex F or m ed in th e Oxid a tive Ad d ition of
P h I to th e P d 0 Com p lex Gen er a ted in Situ fr om
P d (OAc)2 a n d 2 equ iv of d p p p in DMF . Addition of
In contrast to PPh3, for which PhPd(OAc)(PPh3)2 was
formed in the oxidative addition of PhI (Scheme 2),1-3
PhPd(OAc)(dppp) was not formed. It was however
produced quantitatively after addition of nBu4NOAc in
excess (10 equiv) (eq 10) and identified by its 31P NMR
spectrumstwo doublets at -3.7 ppm (J PP ) 48 Hz) and
(14) The conductivity increase is compatible with that observed for
cationic arylpalladium(II) complexes.15
(13) (a) Fitton, P.; Rick, E. A. J . Organomet. Chem. 1971, 28, 287.
(b) Amatore, C.; Broeker, G.; J utand, A.; Khalil, F. J . Am. Chem. Soc.
1997, 119, 5176.
(15) J utand, A.; Mosleh, A. Organometallics 1995, 14, 1810.
(16) Hallberg, A; Larhed, M. Personal communication at the Post-
OMCOS-10 Symposium, Paris, J uly 1999.

Formation of Palladium(0) Complexes
Organometallics, Vol. 20, No. 15, 2001 3245
19.4 ppm (J PP ) 48 Hz)ssimilar to the one reported in
(Scheme 5), has been investigated. They could indeed
coordinate the Pd0 and consequently affect its reactivity
in the course of the oxidative addition.2,3,19 No effect of
I- (5 equiv of nBu4NI added before PhI) was observed
on the kinetics. However, the kinetics was affected by
the presence of AcO- ions. First of all, the oxidative
addition was faster as soon as 1 equiv of nBu4NOAc was
added before PhI. Second, the kinetics became well
reproducible20 in the presence of AcO- with a kinetic
curve still not compatible with an exponential decay
(Figure 3b). This suggests that AcO- ions, when in
excess, may interfere in the oxidative addition process
by changing the structure of the Pd0 complex initially
produced from Pd(OAc)2 and 2 equiv of dppp. When
AcO- ions (10 equiv) were added together with Pd(OAc)2
and 2 equiv of dppp at 48 °C, the rate of formation of
the Pd0 complex was faster but the resulting Pd0
complex was not so stable (the solution turned progres-
sively to brown) and decomposed slowly within 75 min,
after a variable induction period, thus exceeding the
duration required for the investigated oxidative addi-
tions (20 min). The accelerating effect of AcO- ions on
the rate of formation of the Pd0 complex may be
regarded as featuring an irreversible trapping of the
intermediate complex 2 as complex 5 (eq 12), this
reaction being faster than the reactions 2 f 1 or 2 f 3
in Scheme 4. At this level, we do not know whether the
observed Pd0 decomposition is that of complex 5 or of a
further complex.
the literature.17
The addition of nBu4NI in excess (10 equiv) resulted
in the formation of PhPdI(dppp) (eq 11), identified by
comparison of its 31P NMR spectrum with that of an
authentic sample (two doublets at -8.6 ppm (J PP ) 53
Hz) and 12.4 ppm (J PP ) 53 Hz)).13b However this
reaction was considerably slower than that of AcO- ions.
Consequently the structure of the (aryl)PdII(dppp)
complex formed in the oxidative addition of a Pd0
complex generated from Pd(OAc)2/2dppp strongly de-
pends on the aryl derivative considered: (i) four-ligated
cationic ArPd(dppp)(dppp(O))+ complexes are formed
when starting from aryl triflates;16 (ii) cationic ArPd-
(dppp)(dppp(O))+ complexes are formed in the very first
catalytic cycles when starting from aryl iodides and then
neutral ArPdI(dppp) complexes produced by reaction of
I- released in the catalytic reaction; (iii) ArPd(OAc)-
(dppp) complexes are formed in the presence of excess
acetate ions (often used as a base in Heck reactions)5
whatever the aryl iodides/triflates derivatives. We must
also take into account that all these complexes might
be involved in equilibrium, as we previously established
+ 18
for ArPdI(PPh3)2, ArPd(OAc)(PPh3)2, and ArPd(PPh3)2
.
It is worthwhile to note that, in DMF, complexes ligated
by three phosphines, ArPd(PPh3)3+, have never been
detected in appreciable amount even in the presence of
PPh3 in large excess,15 whereas ArPd(dppp)(dppp(O))+
complexes are formed due to less steric hindrance of
dppp when compared to PPh3.
Ra te a n d Mech a n ism of th e Oxid a tive Ad d ition
of P h I to th e P d 0 Com p lex Gen er a ted in Situ fr om
P d (OAc)2 a n d 2 equ iv of d p p p in DMF . The oxida-
tive addition of PhI to the Pd0 complex generated in situ
from Pd(OAc)2 (2 mM in DMF) and 2 equiv of dppp, in
the presence of NEt3 (30 equiv) and H2O (20 equiv), was
monitored by amperometry, as performed previously for
PPh3.2 The rotating electrode was polarized at -0.1 V,
on the plateau of the oxidation wave of the Pd0 complex.
The decay of the oxidation current intensity, propor-
tional to the Pd0 concentration, was recorded versus
time after addition of PhI (5 equiv) at 48 °C. The
solution turned from orange-yellow to pale yellow. From
the curve obtained in Figure 3a, one already observes
that the decay of the Pd0 concentration does not obey a
classical decreasing exponential curve, fitting with a
reaction order of +1 for Pd0, as reported for Pd0(PPh3)2-
(OAc)-,2 suggesting a more complicated mechanism for
the oxidative addition. It is important to note that,
under similar conditions, the Pd0 complex generated
from Pd(OAc)2 and 2 equiv of dppp is considerably less
reactive with PhI (by a factor 300) than Pd0(PPh3)2(OAc)-
generated from Pd(OAc)2 and 3 equiv of PPh3.3
When the AcO- ions were added once the Pd0 complex
has been formed quantitatively from Pd(OAc)2 + 2 equiv
dppp, the decomposition time of the resulting Pd0 was
considerably longer than the time required for the
oxidative addition. This latter procedure was therefore
adopted for the detailed mechanistic investigation of the
oxidative addition. PhI was added just after addition of
AcO- ions. The final complex formed in the oxidative
addition was then PhPd(OAc)(dppp).
The kinetics of the oxidative addition of PhI (5 equiv)
has been investigated as a function of the AcO- con-
centration in the range 1-15 equiv. Figure 4 exhibits
the plot of 1/t1/2 versus AcO- concentration (t1/2 ) half
reaction time of the oxidative addition). The reaction is
then first order in AcO- at low AcO- concentrations,
while it becomes zero order at higher AcO- concentra-
tions (over 7 equiv of AcO-). The saturating effect
suggests that AcO- ions are involved in an equilibrium,
which delivers the reactive Pd0 complex and which is
completely shifted toward this reactive species as soon
as 7 equiv of AcO- are present.
Whatever the AcO- concentration, at low PhI con-
centrations (less than 10 equiv) the reaction order in
Pd0 is neither 1 nor 0 but 1/2, as shown by the linear
In a first approach, the influence of ions such as AcO-
or I-, which are released in the oxidative addition
(17) Ludwig, M.; Stro¨mberg, S.; Swensson, B.; Akermark, B. Or-
ganometallics 1999, 18, 970.
(18) Amatore, C.; Carre´, E.; J utand, A. Acta Chem. Scand. 1998,
52, 100.
(19) Amatore, C.; Azzabi, M.; J utand, A. J . Am. Chem. Soc. 1991,
113, 8375.
(20) The kinetics were not reproducible in the absence of added
AcO-, probably because of the formation of aggregates.

3246 Organometallics, Vol. 20, No. 15, 2001
Amatore et al.
F igu r e 3. Kinetics of the oxidative addition of PhI to the
Pd0 complex generated from Pd(OAc)2 (2 mM) and dppp (4
mM) in the presence of H2O (40 mM) and NEt3 (60 mM),
monitored as a function of time at a rotating gold disk
electrode (2 mm diameter, ω ) 105 rad s-1) polarized at
-0.1 V. 48 °C. (a) Variation of the oxidation current
intensity i (proportional to the Pd0 concentration) with time
in the presence of PhI (10 mM). (b) Same reaction in the
presence of nBu4NOAc (30 mM) added to the Pd0 complex
before PhI.
F igu r e 5. Kinetics of the oxidative addition of PhI to the
Pd0 complex generated from Pd(OAc)2 (2 mM) and dppp (4
mM) in the presence of H2O (40 mM) and NEt3 (60 mM)
monitored at a rotating disk electrode, as a in Figure 3b,
at 48 °C. (a) nBu4NOAc (8 mM), PhI (10 mM); plot of (i/
i0)1/2 ) ([Pd0][Pd0]0)1/2 vs time (i ) oxidation current
intensity of Pd0 at t; i0 ) initial oxidation current intensity
of Pd0). (b) nBu4NOAc (30 mM), PhI (160 mM); plot of (i/
i0)1/2 ) ([Pd0][Pd0]0)1/2 vs time. (c) nBu4NOAc (30 mM), PhI
(160 mM); (+): plot of i (i ) oxidation current intensity of
Pd0 at t) vs time. The solid line is the simulated theoretical
curve using the kinetic law: i ) (i0/C0){(C01/2 + kobs/k′obs)-
obs
e-k′ ×t/2 - kobs/k′obs}2 with kobs ) 5.3 × 10-5 M1/2 s-1, k′obs
)
2.1 × 10-3 s-1, and Co ) 2 mM.
of a second kinetics with a reaction order of +1 for Pd0,
as established by comparison with a theoretical kinetic
curve (Figure 5c, [PhI] ) 160 mM, [Pd0]0 ) C0 ) 2 mM)
F igu r e 4. Kinetics of the oxidative addition of PhI to the
Pd0 complex generated from Pd(OAc)2 (2 mM) and dppp (4
mM) in the presence of H2O (40 mM) and NEt3 (60 mM)
monitored as a in Figure 3, at 48 °C. The kinetics is
investigated in the presence of various amounts of AcO-
ions added as nBu4NOAc just before PhI (10 mM). Plot of
1/t1/2 (t1/2 ) half reaction time) vs AcO- concentration (total
concentration including the acetate released from
AcOH when the reaction 7 is performed in the presence of
NEt3).
simulated using the kinetic law of eq 14c, with kobs
)
5.3 × 10-5 M1/2 s-1 and k′obs ) 2.1 × 10-3 s-1
.
d[Pd0]/dt ) -k′obs[Pd0] - kobs[Pd0]1/2
(14a)
obs×t/2
2
[Pd0] ) {([Pd0]01/2 + kobs/k′obs)e-k′
- kobs/k′obs
}
(14b)
plot of (i/i0)1/2 ) ([Pd0][Pd0]0)1/2 versus time (Figure 5a)
(i ) oxidation current intensity of Pd0 at t; i0 ) initial
oxidation current intensity of Pd0), in agreement with
the kinetic law given in eq 13.
i ) (i0/C0){(C0 + kobs/k′obs)e-k′
- kobs/k′obs
}
1/2
obs×t/2
2
(14c)
According to this hypothesis, the oxidative addition
proceeds by two parallel pathways, one with a reaction
order of +1/2 in Pd0, the second one involving a reaction
order of +1 in Pd0 showing up mostly at high PhI
concentrations. This dual mechanism is operating what-
ever the AcO- concentration. The reaction order in PhI
was determined at high AcO- concentrations (>7 equiv)
when the kinetics did not depend on the AcO- concen-
trations. The plots of kobs and k′obs versus PhI concentra-
tion are reported in Figures 6a,b, respectively. When
the reaction order in Pd0 is 1/2, the reaction order in
d[Pd0]/dt ) -kobs[Pd0]1/2
([Pd0]/[Pd0]0)1/2
)
-kobst/2[Pd0]0 + 1 ) (i/i0)1/2 (13)
1/2
kobs is then calculated from the slope (Figure 5a): kobs
) 8.6 × 10-5 M1/2 s-1 (vide infra).
At higher PhI concentrations (over 10 equiv) and
whatever the AcO- concentration, we observed a devia-
tion from the reaction order of 1/2 (Figure 5b) at longer
reaction times. This is consistent with the superposition

Formation of Palladium(0) Complexes
Organometallics, Vol. 20, No. 15, 2001 3247
Sch em e 7. Mech a n ism for th e Oxid a tive Ad d ition
This kinetic law is then in agreement with the experi-
mental reaction order of 1/2 found for Pd0 and 0 for PhI
with kobs ) (5 ( 1) × 10-5 M1/2 s-1 (Figure 6a), a value
compatible with that found above.
1/2
kobs ) 2k2(k1/k-1
)
1/2, then k2(k1/k-1
)
)
(2.5 ( 0.5) × 10-5 M1/2 s-1
• At higher PhI concentrations, k0[PhI] . k1 and
the oxidative addition proceeds not only from A′ but
also from A2 in competition with its dissociation (Scheme
6).
The kinetic law is then expressed as in eq 16a.
v ) k0[A2][PhI] + 2k3[A′][PhI] ) k0[A2][PhI] +
2k2k3[A][PhI]/(k-2 + k3[PhI]) (16a)
F igu r e 6. Kinetics of the oxidative addition of PhI (in the
range 20-350 mM) to the Pd0 complex generated from
Pd(OAc)2 (2 mM) and dppp (4 mM) in the presence of H2O
(40 mM) and NEt3 (60 mM) monitored at a rotating disk
electrode, as a in Figure 5c, at 48 °C. v ) k′obs[Pd0] +
With the same hypothesis as above: k3[PhI] . k-2, one
gets eq 16b.
k
obs[Pd0]1/2. (a) Variation of kobs vs PhI concentration. (b)
Variation of k′obs vs PhI concentration.
v ) k0[A2][PhI] + 2k2(k1/k-1)
1/2[A2]1/2 (16b)
Sch em e 6. Min im a l Mech a n ism for th e Oxid a tive
Ad d ition
The kinetic law (eq 16b) is in agreement with the
experimental reaction order of 1/2 for Pd0 and 0 for PhI
found above in addition to an experimental reaction
order of 1 found for Pd0. However, in the latter case,
the reaction order is strictly 1 in PhI, which was not
observed experimentally (see the variation of k′obs in
Figure 6b). This implies that the dimer A2 should also
disappear along a concurrent path being zero order in
PhI. This may arise to give an inactive or active dimeric
complex A2* (Scheme 7).
PhI is 0 (kobs is constant, Figure 6a). This means that
the rate-determining step of the overall oxidative ad-
dition does not involve PhI but is a step in which a
reactive monomeric Pd0 complex is formed from a dimer
(order 1/2 in Pd0). When the reaction order in Pd0 is +1,
the reaction order in PhI is not strictly +1 since the
straight line for the variation of k′obs does not intercept
at the origin (Figure 6b); thus k′obs ) a[PhI] + b. On
the basis of these experimental results, the simplest
mechanism, which rationalizes all findings, is proposed
in Scheme 6. A reactive dimeric Pd0 species A2 leads to
two successive equilibria involving a nonreactive mono-
meric Pd0 species A and another one A′, which reacts
with PhI in parallel (Scheme 6).
Indeed, provided that the various rate constants are
adequate, this mechanism displays all the experimen-
tally observed features:
• At low PhI concentrations, k0[PhI] , k1 and the
monomer A′ is the only reactive species. The kinetic law
for the disappearance of the dimer A2 is then given by
the eq 15a, after considering a fast equilibrium between
A2 and 2A and the steady-state approximation for A′.
The corresponding kinetic law is then given in eq 17.
v ) k0[A2][PhI] + k4[A2] + 2k2(k1/k-1
k′obs[A2] + kobs[A2]1/2 (17)
k′obs ) k0[PhI] + k4 ) a[PhI] + b
) )
1/2[A2]1/2
(see eq 14a and Figure 6b)
From Figure 6b one deduces k0 ) 6.5 × 10-3 M-1 s-1
and k4 ) 1.1 × 10-3 s-1
.
The saturating effect of acetate ions on the kinetics
of the oxidative addition, observed at high AcO- and
low PhI concentrations (Figure 4), implies that the AcO-
ions are involved in an equilibrium, which produces the
Pd0 dimer A2 with a reaction order of +1 for AcO-
(Figure 4). This equilibrium would be completely shifted
toward the dimer A2 as soon as 7 equiv of AcO- are
added to the solution, in which a Pd0 complex has been
formed spontaneously from Pd(OAc)2 and 2 equiv of
dppp. The complex A2 is then the major Pd0 species.
This suggests that the Pd0 complex, formed from
Pd(OAc)2 and 2 equiv of dppp, is not Pd0(dppp)(OAc)-
(3) as tentatively proposed in Scheme 4, but probably a
v ) 2k3[A′][PhI] ) 2k2k3[A][PhI]/(k-2 + k3[PhI])
(15a)
When k3[PhI] . k-2, one obtains eq 15b.
dimer A2 formed from Pd0(dppp)(OAc)-, due to the
0
v ) 2k2[A] )2k2(k1/k-1)
1/2[A2]1/2 ) kobs[A2]1/2 (15b)
0
bidentate character of dppp (Scheme 8). The dimer A2 ,

3248 Organometallics, Vol. 20, No. 15, 2001
Amatore et al.
Sch em e 8. Mech a n ism for th e Oxid a tive Ad d ition of P h I to th e P d 0 Com p lex Gen er a ted fr om P d (OAc)2
a n d 2 equ iv of d p p p in th e P r esen ce of AcO- Ion sa
aFor more clarity, the two Ph ligands of the P atoms have been omitted.
with two 16-electron Pd0 centers, can be coordinated to
extra AcO- to form a new Pd0 dimer, A2. This dissociates
into a nonreactive 18-electron complex A. The cleavage
of the a P-Pd bond is then required to produce a
reactive 16-electron Pd0 complex, A′ (Scheme 8). At high
PhI concentration, the dimer A2 is found to be reactive,
although each Pd0 center has 18 electrons. It probably
reacts after an equilibrated cleavage of one P-Pd bond
without any modification of the kinetic law. k0 would
then be the apparent rate constant of the overall
reaction, i.e. fast equilibrated cleavage of one P-Pd
bond (equilibrium constant K0) followed by a slow
oxidative addition of PhI (rate constant k′0) with k0 )
K0k′0.
its decomposition by protons, generated by the hydroly-
sis of a phosphonium salt (Scheme 4). In the presence
of 3 equiv of dppp, a nonreactive Pd0(dppp)2 complex is
generated. The oxidative addition with PhI gives a
cationic complex PhPd(dppp)(dppp(O))+ (in which
dppp(O) behaves as a monodentate ligand) or PhPd-
(OAc)(dppp) in the presence of added AcO-. The oxida-
tive addition of PhI is an intricate reaction whose
kinetics could be fully investigated only in the presence
of added AcO-. Its involves reactive dimeric or/and
monomeric Pd0 complexes ligated by AcO-, whose
respective reactivity is a function of the PhI concentra-
tion (Scheme 8). The complexity of the oxidative addition
pathways when performed with a dppp ligand compared
to that with a PPh3 ligand stems from the bidentate
character of the dppp ligand which favors dimeric
structures.
We are conscious that none of the proposed structures
0
for complex 3, A2 , and A2 could be characterized. We
only know the 31P NMR shift of the Pd0 complex formed
from Pd(OAc)2/2dppp in the absence of extra AcO- at
0
4.2 ppm, which is assigned to complex A2 . However,
Exp er im en ta l Section
the proposed structures for complexes A2, A, and A′ are
consistent with the experimental kinetic results and
kinetic laws. So we use them here only as a chemical
rationale of the complex mechanistic sequence, which
has been fully characterized kinetically (Scheme 7). We
do not propose any structure for A2*, although there
Gen er a l Rem a r k s. 31P NMR spectra were recorded on a
Bruker spectrometer (101 MHz) in DMF containing 10% of
acetone d6. Cyclic voltammetry was performed with a home-
made potentiostat and a waveform generator Tacussel GSTP4.
The cyclic voltammograms were recorded on a Nicolet 301
oscilloscope. Conductivities were measured on a Tacussel
CDM210 conductimeter (cell constant ) 1 cm-1).
0
might be an identity between A2* and A2 .
Ch em ica ls. DMF was distilled from calcium hydride under
vacuum and kept under argon. Pd(OAc)2, dppp, PhI, nBu4NI,
and nBu4NOAc were commercial.
Con clu sion
As for monodentate phosphine, a Pd0 complex is
spontaneously formed from Pd(OAc)2 and a bidentate
phosphine such as dppp. dppp is the reducing agent and
is oxidized to the hemioxide dppp(O). In contrast to
PPh3, the intramolecular reduction step is reversible.
Consequently, a stable Pd0 complex is quantitatively
formed in the presence of 2 equiv of dppp, water, and a
base (NEt3), which stabilized the Pd0 complex vis a` vis
Electr och em ica l Setu p a n d Electr och em ica l P r oce-
d u r e for Volta m m etr y. Experiments were carried out in a
three-electrode thermostated cell connected to a Schlenk line.
The counter electrode was a platinum wire of ca. 1 cm2
apparent surface area. The reference was a saturated calomel
electrode (Tacussel) separated from the solution by a bridge
filled with 3 mL of DMF containing nBu4NBF4 (0.3 M). DMF
(12 mL) containing nBu4NBF4 (0.3 M) was introduced into the

Formation of Palladium(0) Complexes
Organometallics, Vol. 20, No. 15, 2001 3249
cell followed by 5.4 mg (0.024 mmol) of Pd(OAc)2 and then 1-3
equiv of dppp. Cyclic voltammetry was performed at a steady
gold disk electrode (0.5 mm or 2 mm diameter) at a scan rate
of 0.2 V s-1. Kinetic measurements for the Pd0 formation and
its reactivity with PhI were performed at a rotating gold disk
electrode (2 mm diameter, ω ) 105 rad s-1) polarized at -0.1
vs SCE.
Ack n ow led gm en t. This work has been supported
in part by the Centre National de la Recherche Scien-
tifique (CNRS, UMR 8640) and the Ministe`re de la
Recherche (Ecole Normale Supe´rieure).
OM0101137