Y. Murata et al. / Tetrahedron Letters xxx (xxxx) xxx
3
Table 2
screened (entries 1, 8–13), keeping the Pd catalyst and other
Substrate scope reaction of triarylbismuthines with azole-2-thiones.a,b
parameters of standard conditions (entry 1) unchanged. Among
all the Cu reagents, Cu(OAc)2 was found to be the most effective
one. CuTC is reported to be a widely used copper reagent in Liebe-
skind-Srogl reactions [4–7,10]. However, we observed that Cu
(OAc)2, which is more easily available, showed much better results
in case of the reactions, we carried out, using bismuth reagent. The
results of solvent screening showed that the reaction between 1a
and 2a, using Pd(dba)2 and Cu(OAc)2, proceeded effectively in
DMF (86%), NMP (83%), toluene (79%), 1,4-dioxane (79%), and
DMSO (73%) (entries 1, 15–18), whereas MeCN (16%), and EtOH
(16%) exhibited inferior results (entries 1, 19–21).
Next, coupling of 1a was performed with various aryl donors
(2a, 3–7), containing bismuth and antimony, in DMF at 80 °C, using
Pd(dba)2 catalyzed and Cu(OAc)2 mediated conditions (entries 1,
22–26). The reaction proceeded smoothly with the trivalent
organobismuth compound (2a) (entry 1), whereas the use of pen-
tavalent Ph3BiCl2 (3) and Ph3Bi(OAc)2 (4) not only decreased the
yield of 8 but also necessitated extended reaction times (entries
22–23). Moreover, all antimony reagents such as Ph3Sb (5), Ph3-
SbCl2 (6), and Ph3Sb(OAc)2 (7) failed to undergo coupling with
1a. These results indicate that the reactivity of triphenylpnictogen
reagents are considerably affected by the constitutive metal (Bi, Sb)
as well as their valencies [(III), (V)].
We also investigated the effect of reducing reaction tempera-
ture by performing the coupling reaction between 1a and 2a at
60 °C and keeping all other parameters of standard conditions
unchanged (entry 27). In this case, although 8 was formed in 54%
yield, reaction time increase to 24 hr. A significant decrease in
the yield of product 8 and increase in that of byproduct 9 were
observed when coupling was carried out under O2 and air instead
of Ar (entries 28, 29). Decreasing the catalyst [Pd(dba)2] loading
from 10 to 5.0 mol% resulted in decreased yield of 8 and extended
the reaction time (entry 30). Similar observations were also made
when the catalytic amount of Cu(OAc)2 (30 mol%) was used instead
of 2.0 equiv., the standard amount (entry 31).
After screening various reagents and changing reaction condi-
tions, we observed that the best result was obtained when 1a
was treated with Ph3Bi (2a) in the presence of Pd(dba)2 (10 mol
%) and Cu(OAc)2 (2.0 equiv.) in DMF at 80 °C under argon atmo-
sphere. As Ph3Bi (2a) has three phenyl groups, the coupling reac-
tion was performed using a 3: 1 molar ratio of benzothiazole-2-
thione (1a: 0.75 mmol) and Ph3Bi (2a: 0.25 mmol). However, the
yield was low (29%), indicating that only one of the three phenyl
groups on bismuth is involved in C(HetAr)–C(Ar) bond formation
under the present reaction conditions (entry 32). In Pd-catalyzed
cross-coupling reactions using triarylbismuthine, bases are widely
used for the activation of bismuth reagents [17–25]. However,
when we performed coupling of 1a and 2a using Cs2CO3 as a base,
we did not observe any yield improvement (entry 33).
In order to investigate the efficiency and substrate scope of the
Liebeskind-Srogl-type reaction described in this work, we per-
formed reactions of various azole-2-thiones (1a-l) (0.5 mmol) with
Ar3Bi (2a-k) (0.5 mmol) under the previously optimized conditions
(Table 2). Reactions of 1a with various Ar3Bi (2b-g) proceeded suc-
cessfully to achieve C(HetAr)–C(Ar) bond formation. However, the
yield of the coupling products was observed to be dependent on
the electronic nature of the substituents on the phenyl rings of Ar3-
Bi. For instance, Ar3Bi bearing a methyl group, halogens, and elec-
tron-withdrawing groups such as ethoxycarbonyl and
trifluoromethyl on the phenyl rings produced coupling products
(11–15) in good-to-high yield. In contrast, Ar3Bi, containing strong
electron-donating groups, such as methoxy, produced coupling
product 10 in low yield. Sterically hindered ortho-substituted Ar3-
Bi (2 h-j) was coupled with 1a, affording the respective coupling
products (16–18, respectively) with relatively lower yield, except
a1a-l (0.5 mmol), 2a-k (0.5 mmol), Pd(dba)2 (0.05 mmol), Cu(OAc)2 (1.0 mmol).
bIsolated yield based on one aryl group in triarylbismuthines.
for the ortho-methyl derivative (16). The reaction of 1a with 2k
having a thienyl group generated 19 in lower yield. Next, the reac-
tion of various benzothiazole-2-thiones (1b-d), bearing electron-
donating or electron-withdrawing groups at the 5-position, was
performed with 2a, affording the corresponding coupling products
(20–22) in high-to-excellent yield. Additionally, the reaction of
various other benzoxazole-2-thiones (1e-h) with 2a generated
the corresponding oxazoles (23–26, respectively) without any dif-
ficulty. Similarly, the reaction of thiazolopyridine-2-thione (1i) and
2a afforded 27, but in moderate yield. Moreover, the reaction of N-
methylbenzimidazole-2-thione (1j) with 2a produced 28 in 51%
yield. On the other hand, the reaction using a benzimidazole-2-
thione (1k) instead of 1j yielded a complex mixture, and the
excepted coupling product (29) was not obtained. Monocyclic thi-
azole-2-thione (1l) reacted with 2a to afford the desired product 30
in 56% yield.
At present, the reaction mechanism of the C(HetAr)–C(Ar) bond
formation, reported in this paper, is unclear. We presume that the
Please cite this article as: Y. Murata, A. Terazoe, M. Kiba et al., Liebeskind-Srogl-type cross-coupling reaction of azole-2-thiones with triarylbismuthines: