L.-C. Liang et al. / Journal of Organometallic Chemistry 804 (2016) 30e34
31
Fig. 1. Representative examples of phosphine palladacycles.
previously established Heck [8a] and Sonogashira reactions [8f].
This study also represents the first report of Suzuki couplings
catalyzed by PNP complexes, thus effectively widening the appli-
cations of these complexes on catalysis.
homogeneous instead of a result from colloidal or bulk palla-
dium(0) [14]. Given the increased demand on the development of
green and sustainable syntheses, water-compatible homogeneous
catalysts are attractive and valuable as water is the most environ-
mentally benign, abundant, and safe solvent [11e,15].
To identify the reactivity preferences of carbon-halide bonds in
this catalysis, fluorinated (entries 14e15) and iodinated (entry 34)
phenyl bromides were examined, from which fluoro and bromo
biaryls, respectively, were selectively produced as the cross-
coupling products. The reactivity of aryl halides thus follows the
order I > Br > F. In contrast to 3 and 4 [8b], 5a is rather inactive
toward chloro substrates (entries 36e37), again reflective of the
significance of differences in catalyst compositions.
2. Results and discussion
Compounds 5a and 5b were prepared following literature pro-
cedures [8a,8f]. Both complexes are not sensitive to air or moisture
and remain intact upon heating in solutions at 110 ꢀC or higher. To
survey reaction parameters for Suzuki couplings, we chose to
examine the reactivity of 5a toward the reaction of 4-tolyl bromide
with phenylboronic acid in the presence of a variety of bases
(K3PO4$H2O, K3PO4, K2CO3, Cs2CO3, Na2CO3, NaHCO3, NaOtBu) and
solvents (dioxane, toluene, DME) under aerobic conditions. Due to
the inherently low solubility of these bases in selected solvents at
room temperature, heating the reaction mixtures to 80 ꢀC or higher
was attempted to facilitate this catalysis. In general, reactions run
with K3PO4$H2O or K3PO4 in dioxane or toluene solutions outper-
form the others, producing 4-methylbiphenyl as the desired
product in satisfactory yields (Table 1, entries 1e5).
Reaction scopes of this catalysis were examined with conditions
resolved from the parameter survey. As summarized in Table 1, a
number of electronically activated, unactivated, and deactivated
aryl bromides (entries 1e32) and iodides (entries 33e35) are
competent building blocks. Compatible functional groups include
keto, aldehyde, nitro, fluoro, chloro, alkyl, alkoxy, and amino, etc.
Satisfactory reaction yields were realized with the employment of
0.1 mol% of 5a under aerobic conditions or even in the presence of
exogenous water (entry 7). The development of water-compatible
Suzuki coupling catalysis is of current interest [11]. Lowering the
catalyst loading to 0.01 mol% (entries 6 and 28) gives higher
turnover numbers of up to 6.2 ꢁ 103 based on the substrates that
were surveyed. This reaction recipe also allows for successful
couplings of heterocyclic (entry 16) and 2,6-disubstituted (entries
20e26) substrates. Of particular note is the high yield production of
tri-ortho-substituted biaryls (entries 25e26). Remarkably, 5a gave
the sterically encumbered 2,4,6-triisopropyl-20-methylbiphenyl
[12] in a much higher yield than 3 by 38% under similar conditions
[8b], highlighting a significant improvement in the synthesis of
sterically hindered biaryls with catalytic amido phosphine com-
plexes of palladium. This apparent discrepancy in activities of 5a
and 3 also underscores the significant impacts of a seemingly minor
change in catalyst compositions on catalysis scopes.
In principle, a more electron-rich and sterically demanding
palladium complex would electronically encourage oxidative
addition of aryl halides and sterically facilitate reductive elimina-
tion of CeC bond forming products. In view of this, 5b that is
constitutionally more electron-releasing and more sterically
demanding due to P-isopropyl [8f] groups incorporated was subject
to catalysis examinations (entries 38e43), but it turned out to be
consistently much less reactive. The reaction yields are all signifi-
cantly lower than those in the corresponding reactions employing
catalytic 5a under similar conditions, thus highlighting a profound
P-substituent effect of these amido phosphine complexes. Complex
5a that is less sterically demanding and less electron-releasing, on
the other hand, should sterically advance oxidative addition and
electronically expedite reductive elimination. It has been shown
that nickel analogues of 5a have a much higher inclination to
reductive elimination than those of 5b [8n]. In addition, an increase
in the coordination number of these palladium catalysts after
oxidative addition is highly anticipated given the fact that the
phosphorus donors in analogous group 10 complexes do not tend to
dissociate readily [8c,8h,8k,16]. As a result, the less sterically
demanding 5a should also be advantageous in oxidative addition.
To acquire more mechanistic evidences, competitive experi-
ments were performed employing catalytic 5a for couplings of
activated, unactivated, and deactivated aryl bromides with phe-
nylboronic acid in either dioxane at 110 ꢀC or toluene at 100 ꢀC,
leading to Hammett plots (Fig. 2) with reaction constants
r of 0.25
and 1.08, respectively. These small values also suggest that the
r
oxidative addition of aryl bromides is not rate-limiting in this
catalysis. A similar conclusion was also deduced from Hammett
plots of 3- or 4-catalyzed Suzuki couplings (
respectively) [8b] and 5a-catalyzed Heck olefination (
or Sonogashira alkynylation (
¼ 0.82) [8f]. In contrast to these
small values, larger reaction constants [17] have been consistently
reported for oxidative addition of aryl halides to Pd(0), e.g.,
¼ 2.3
for aryl chlorides to Pd(XPhos) [17a] and
¼ 2 for aryl iodides to
r
¼ 0.48 and 0.66,
r
¼ 0.60) [8a]
r
The ability to achieve this catalysis at elevated temperatures in
the presence of water under aerobic conditions is remarkable,
considering the conformation of 5a that is composed of an intrin-
sically highly basic amide and a highly reactive PdeN bond [13]. No
palladium black was observed in all reactions shown in Table 1. To
probe its homogeneity, entry 2 was repeated with addition of an
equal volume of water to toluene and 1360 equivalents of mercury
to 5a. The reaction yield of this controlled experiment was essen-
tially unaffected, thus indicating that this catalysis is indeed
r
r
r
Pd(PPh3)2 [17b]. In view of the facile reductive elimination rates
found for nickel analogues of 5a [8n], we propose that reductive
elimination of CeC bond forming products is not rate-limiting,
either. Thus, transmetallation is more likely to be the slowest in
the current study. Similar conclusions were also made in other
Liang, Lan-Chang
