G Model
CCLET 5587 No. of Pages 5
2
Z. Yao et al. / Chinese Chemical Letters xxx (2019) xxx–xxx
Ph2
P
Ph2
P
Ph2
P
CH2R
CHR
C X
PPh2
PhosphYl (non salt-freeylide)
CHR
[M]
PPh2
PPh2
RCH3X
[MY]
B C
–BH
X
Y
+
C X
PPh2
phosphino-phosphonium salt
P
Ph
Ph
PhosphYl complex
C = Li, Na, K, Cs...
R=H or CO2Et
Ph2
P
Ph2
P
PPh2
X
CH2R
CH2R
X
X = I, TfO, PF6
PPh2
"dppbzium"
L1•HX
"BINAPIUM"
L2•HX
Scheme 1. Phosphine-phosphonium salts and ylides there of (PhosphYls) as LX-chelating ligands of late transition metals M (e.g., Ru, Rh, Pd).
Scheme 1) used in Pd-catalysedC-allylation[14] (nottobe confused
with YPhos, ylido-phosphines acting as strongly donating P ligands
in Pd or Au catalysts) [15]. As the low pKa values of Yliphos favours
partial de-coordination of the stabilized free ylide [17], exploratory
investigations prompted the choice of the rac-methyl-BINAPIUM
ylide L2 as a benchmark PhosphYl. On the tracks of pioneering
works [16] and recent advances in Pd-catalysed CꢀꢀH arylation of
heterocycles [17], the transformation was chosen as a model to
further evaluate the ability of PhosphYls to form in situ catalysts.
Indeed, while their electron-donating ability is expected to
facilitate the step of the aryl halide oxidative addition, their bulky
chelating characters appear appropriate for steric protection of the
Pd center.
backbone. The use of the BINAPIUM ylide generated from L2ꢂHI
gave 3a in an improved yield of 65% (entry 12). The same ylide L2
generated from L2ꢂHOTf gave 3a in a much higher yield of 93%,
revealing a dramatic counterion effect (entry 13). This was
confirmed by the result of L2ꢄHPF6, giving 3a in a twice lower
yield of 50% (entry 14), thus showing a correlation of the yield with
the association effect between the anion and the Pd and/or P+
centers [27].
A systematic screening of the reaction conditions was then
undertaken (Table 1). In the absence of any potential P-ligand, the
reactionwas found to proceed with 16% yield only (entry 15). It was
also found that both the solvent and base are of critical importance.
By reducing the amounts of Cs2CO3 from 1.0 equiv. to 0, the yields
in 3a are reduced to 41%, 33% and 0% respectively (entries 16–18).
Upon replacement of Cs2CO3 by weaker bases, such as Et3N or PPh3,
product was not observed (entries 19 and 20). The use of a stronger
base, such as NaHMDS, tBuOK or MeONa, gave lower yields in the
range of 27%–39% (entries 21–23). When the same bases were used
in THF instead of DMSO, the yields dropped to 12%–20% (entries
24–26). Other solvents, such as THF, DMF, MeCN, EtOH or NMP, also
resulted in a dramatic decrease of the yield (< 15%, details see the
Supporting information).
With the view to accessing functional hetero-aromatic systems
without resorting to an ortho-directing group, the reaction of the
benzoxazole substrates
1 with various aryl halides 2 was
considered [7]. The target 2-aryl-benzoxazole products 3 indeed
exhibit versatile properties [18], going from antibacterial [19],
cholesterol ester transfer protein inhibition [20], adenosine
receptor antagonism [21], amyloidogenesis inhibition [22], and
thioflavine-like binding affinity to β-amyloid plaques [23], to
fluorescence in the solid state [24].
As a soluble Pd(II) salt widely used in CꢀꢀH activation catalysis
also exhibiting propensity to form Pd–C(sp3) bonds [25], Pd(OAc)2
was envisaged as catalyst precursor, to which non-salt-free
PhosphYls could be added, after in situ deprotonation of parent
phosphoniums. Considering the possible effect of the anion Xꢀ in
non-salt-free ylides, well known in the Wittig reaction [26], several
The yield in 3a was decreased from 93% (entry 13) with the
salted ylide (L2/LiOTf) to 46% (entry 27) with the salt-free ylide,
similar to the 50% yield (entry 14) obtained with the pseudo-salt-
free ylide L2/LiPF6 (contrary to TfOꢀ, PF6 is totally inert with
ꢀ
respect to the Pd and P+ centers). Under the same conditions, 1
equiv. of LiOTf was added to the salt-free ylide L2, the yield in 3a
was restored to 80%, thus demonstrating the key role of LiOTf
(entry 28). The effect of LiOTf could be attributed to the interaction
ability of the TfO– anion with the P+ and/or the Pd(II) centers, by
either electrostatic contact or coordination bonding [27].
With the optimized reaction conditions in hand, the scope of
the substrates was examined (Scheme 2). Benzoxazole 1 reacted
smoothly with bromobenzene 2a and its derivatives 2b-2t to give
the anticipated products 3a-3t in moderate to high yields
(35%–93%). Aryl bromides with a methyl group at the para- or
meta-position (2c-2d) gave the corresponding products 3c and 3d
in 82% and 79% yields, respectively. When the methyl group was
attached at the ortho-position, a lower 54% yield was obtained,
revealing the steric hindrance of the reacting position. The
electron-withdrawing substituent CF3 at the para-position of 2e
allowed isolation of the aryl-benzoxazole 3e in a slightly higher
yield (84%), while electron-donating OMe counterpart has the
opposite effect on the corresponding product 3f (70%). Neverthe-
less, both weakly electron-withdrawing (F, Cl) and strongly
electron-donating (NMe2, tBu) substituents gave higher yields
phosphonium salts L2ꢂHX were compared (X
¼
I, TfO, PF6).
Benzoxazole 1 and bromobenzene 2a, or 4-bromoanisole 2b,
were selected as the model reactants. From preliminary
a
screening under various conditions, the following conditions were
adopted: DMSO as a solvent, Cs2CO3 as a base and Pd(OAc)2 as
catalyst precursor at 50 ꢃC for 14 h (Table 1). At the outset,
commercially available monodentate phosphines, PPh3, PCy3, P
tBu3, were first selected as standard reference ligands, giving 3a in
48%, 45% and 37% yields, respectively (entries 1–3). Bidentate
phosphines BINAP, dppm, o-dppbz, dppe and dppf also gave
medium yields of 53%, 44%, 51%, 47% and 36%, respectively (entries
4–8). Under the same conditions, using the methylphosphonium
salt L1ꢂHX (X
I, TfO, PF6) of o-dppbz as pro-ligand (non-salt-free
¼
PhosphYls L1/HX or L2/HX were separately prepared by depro-
tonation of L1ꢂHX or L2ꢂHX with nBuLi in DMSO solution during
15 min at room temperature), the coupling product 3a was isolated
in 53%, 59% and 43% yields, respectively (entries 9–11). These
results are similar to those obtained by bidentate phosphines, and
then it was decided to change the o-dppbz backbone to the BINAP
Please cite this article in press as: Z. Yao, et al., Phosphine-phosphonium ylides as ligands in palladium-catalysed C2-H arylation of
Yao, Zhenyu
