20
S.J. Sabounchei, M. Ahmadi / Inorganica Chimica Acta 405 (2013) 15–23
aReaction conditions for entries 1–13: aryl halide (1 mmol), ethyl acrylate (2.2 mmol), Cs2CO3 (1.5 mmol), NMP (2 ml), catalysts 1 and 2 (0.001 mol%), 130 °C.
bReaction conditions for entries 14–26: aryl halide (1 mmol), styrene (2.2 mmol), K2CO3 (1.5 mmol), DMF (2 ml), catalysts 1 and 2 (0.001 mol%), 130 °C.
cAll reactions carried out under aerobic conditions.
dIsolated yield.
eReaction conditions for entries 27–30: with precursor complex [Pd(dppe)(OTf)2] as catalyst (0.001 mol%). See Ref. [17].
tron-neutral and electron-rich aryl iodides react with olefins to
generate the corresponding cross-coupling products in high yields
(Table 4, entries 1, 4, 14 and 17). For instance, couplings of 4-meth-
yliodobenzene with the olefins proceeded in short times with good
yields in both of the catalysts (Table 4, entries 4 and 17). As shown
in Table 4, activated aryl bromides such as 4-nitrobromobenzene
underwent the Heck reaction with ethyl acrylate and styrene
under similar conditions to afford the corresponding products in
83–85% yields (Table 4, entries 11 and 24) whereas, inactivated
aryl bromides such as 4-methylbromobenzene gave 70–74% yields
(Table 4, entries 5 and 18). The electronically neutral bromoben-
zene (Table 4, entries 2 and 15) produced good amounts of desired
products. Trying to apply 2-bromothiophene and 1-bromonaph-
thalene as efficient substrate was successful (Table 4, entries
12–13 and 25–26). To extend the scope of our work, we next inves-
tigated the coupling of various aryl chlorides with ethyl acrylate
and styrene. A number of reported catalysts in literature need to
be used in high loadings, and showed little or no activity with aryl
chloride substrates [21]. As can be seen in Table 4, aryl chlorides
with electron-withdrawing groups reacted smoothly, and the cor-
responding products were obtained in good yields (Table 4, entries
8, 10, 21 and 23). But for chlorobenzene and 4-methylchloroben-
zene, lower yields were observed even after prolonged reaction
times (Table 4, entries 3, 6, 16 and 19). By using aryl chlorides a
good amounts of stilbene derivatives are yielded under the same
conditions employed for aryl iodides and aryl bromides (see Table
4). The higher C–Cl bond strength compared with C–Br and C–I
bonds disfavors oxidative addition step in catalytic coupling reac-
tions. The ideal substrates for coupling reactions are aryl chlorides,
since they tend to be cheaper and more widely available than their
bromide or iodide analogous. As can be seen in Table 4, we ob-
served that in all cross-coupling reactions, catalysts 1 and 2 had
approximately same catalytic activities and similar yields.
It seems that the precursor complex, [Pd(dppe)(OTf)2] can also
be used to behave as an active resource for producing Pd(0). We
used this complex as main catalyst under identical conditions.
The activity of precursor complex, [Pd(dppe)(OTf)2] with an ‘easy’
set of substrates (Table 4, entries 27 and 28) and a ‘difficult’ set
of substrates (Table 4, entries 29 and 30) was tested. Using this cat-
alyst in the reaction between 4-methyliodobenzene and 4-methyl-
chlorobenzene with styrene, gave 65% and 45% yield, respectively
(Table 4, entries 27 and 29). Therefore, the activities of the cata-
lysts 1 and 2 are found to be superior compared with the complex
[Pd(dppe)(OTf)2].
Likely mechanism for the Heck reaction of the aryl halides with
olefins using palladium complexes as catalysts has been proposed
(Scheme 2) by comparison with the literature reports [22].
In Heck cross-coupling reactions, where a reducing agent can-
not be clearly identified, a catalytic cycle involving Pd(II)/Pd(IV)
complexes [23] was first proposed rather than the traditional one
involving Pd(0)/Pd(II) complexes [24]. The first step of the catalytic
cycle would then be the classical oxidative addition of the aryl ha-
lide with a Pd(0) complex, as in a classical Pd(0)/Pd(II) catalytic cy-
cle [2,10a]. Even in the absence of any identified reducing agent, a
number of authors have nevertheless proposed the reduction of the
palladacycle to a low-ligated Pd(0) complex [22h,I,j,24b,25].
Some of the observations made us to think about a mechanism
being similar to that expressed in standard textbooks [26].
Unfortunately, very little is known about the actual structures
of catalytically active PdLn complexes generated from modern
and highly sophisticated Pd-precursors [27]. In addition, there
have been few attempts to elucidate the composition of the active
species present in the Heck cycle. Published studies have been
limited to the use of electrochemistry [28], anionic detection
(ESI-MS, EXAFS) [29] together with detailed kinetic studies, [30]
NMR detections [23a,24a,31] and catalyst poisoning experiments
[32].
Recently, numbers of cyclopalladated complexes have been re-
ported in which the palladium atom is at the zero oxidation state
[28,33]. Several reports on the existence of highly active soluble
palladium(0) colloids (nanoparticles) have established the function
of palladacycles as plain Pd(0) reservoirs in high temperature
applications like Heck and Suzuki couplings [34]. The reduction
of palladium occurs via insertion into the Pd–C bond followed by
a reductive elimination that provides the 2e to reduce Pd(II) into
Pd(0), but this is generally followed by the formation of Pd(0)
nanoparticles. The fact that metallic Hg does not kill the catalysis
is a good point that no metallic Pd is being produced. Electrochem-
istry is indeed one of the most convenient techniques to detect
Pd(0) complexes that are generated in situ from Pd(II) complexes
[35]. In the other way, we previously showed that Pd(II) in this pal-
ladacycle was reduced to Pd(0) by cyclic voltammetry [17]. In addi-
tion, the 31P NMR spectroscopy show that the activation of
palladacycle phosphine complex 1 produce some kind of active
Pd(0) spices [36]. Palladacycle 1 can be reduced to zerovalent pal-
ladium in DMF whilst dppe dioxide was formed. (See Supplemen-
tary data).
As with palladacycle 1 forms Pd(0) unsaturated 14 electron spe-
cies (Scheme 2) [2,16,17,22a,h–j,g], a highly reactive Pd(0) species
does undergo an oxidative addition with ArX. In the next, olefin in-
serts to catalytic cycle. The resulting complex enters the migratory
insertion step via the neutral route, which is revealed by a huge
acceleration by polar solvents, a common ion effect, and other typ-
ical criteria. The steric and bulky properties of bidentate phosphine
ligand (L) can be facilitating this process [22h,37]. A trans-effect of
a strong chelate phosphorus ligand makes the reductive elimina-
tion of alkene very fast, avoiding to the formation of Z-isomer.
The addition of K2CO3 to remove KHCO3 + KX gives catalytically ac-
tive Pd(0) complex (Scheme 2). In this work, with these aryl ha-
lides, several reactions were performed in the presence of olefins.
In the presence of styrene and ethyl acrylate, the regioselectivity
is generally in favor of the linear isomers (E-isomer) [22h]. Based
on the results, production of the E-isomers in all the cases with
good yields are the great advantages of the presented catalysts
(see Table 4).
Although several catalytic systems have been reported to sup-
port Heck C–C coupling reactions, a homogeneous catalyst of this
type is novel because of its P donor of dppe as ligand and P and
CH of phosphorus ylide environment. The homogeneous nature
of the catalysis was checked by the classical mercury test [38].
Addition of a drop of mercury to the reaction mixture did not affect
the yield of the reaction which suggests that the catalysis is homo-
geneous in nature, since heterogeneous catalysts would form an
amalgam, there by poisoning it.
The comparison data presented in Table 5 show the efficiency of
these new catalysts towards the coupling reactions. From an indus-