4
Y. LI ET AL.
(entry 10), CH3CN (entry 11), CH3NO2 (entry 12), N,N-dimethylformamide (DMF,
entry 14), or ethanol (entry 16). Lowering the reaction temperature to 50 ꢀC led to a
poor yield of xanthen-9-azole 3a (entry 17). Much to our satisfaction, 3a was furnished
in a nearly quantitative yield by using 1.5 equiv of xanthene 1a (entry 18), while an
excellent yield was still achieved with a reduced loading of BPO (entry 19). In most
cases, xanthen-9-one 4 was formed.
Under optimized conditions, the C–N bond forming reaction was further explored
(Table 2). It was found that the electronic nature of the substituents on the azolic ben-
zyl ring plays a poor role in the reaction kinetics. With 5-chloro benzotriazole 2b as the
N–H source, poor N1/N3 selectivity was observed probably due to spin delocalization,
and 5- or 6-chloro-substituted triazole products 3b and 3b0 were delivered in 45% and
36% yields, respectively (entry 2). 5,6-Dimethyl benzotriazole 2c (entry 3), 5-phenyl
tetrazole 2d (entry 4), 1,2,4-triazole 2e (entry 5), pyrazole 2f (entry 6), ethyl pyrazole-4-
carboxylate 2g (entry 7), 5-chloro isatin 2h (entry 8), and 3-acetyl indole 2i (entry 9)
are all suitable N–H coupling partners, and they reacted with xanthene 1a to give corre-
sponding xanthen-9-amines 3c–i in high to quantitative yields. Corresponding C–N
bond forming products 3j–l were delivered in only poor to moderate yields from aceta-
mide 2j (entry 10), benzamide 2k (entry 11), or N-methylbenzamide 2l (entry 12),
probably due to insufficiently stabilized N-radicals. 2-Acetylxanthene 1b (entry 13) and
2-benzoylxanthene 1c (entry 14) are less reactive than xanthene 1a, and their reactions
with benzotriazole 2a led to xanthen-9-azoles 3m,n in high yields. Thioxanthenes 1d–f
could be alternative C(sp3)–H coupling partners. Whereas thioxanthene 1d reacted with
benzotriazole 2a (entry 15) or 5,6-dimethyl benzotriazole 2c (entry 16) to afford C–N
coupling products 3o,p in 97% and 90% yields, respectively, thioxanthen-9-azole 3q
derived from 1,2,4-triazole 2e was produced in a good yield (entry 17). 2-Isopropyl
thioxanthene 1e (entry 18) and 2-chloro thioxanthene 1f (entry 19) are competent
C(sp3)-H sources as well, and corresponding xanthen-9-azoles 3r,s derived from 2a
were furnished in 90% and 92% yields, respectively. The remote activating heteroatom
is essential, and diphenylmethane 1g without such a group reacted with benzotriazole
2a (entry 20) or 5-phenyl tetrazole 2d (entry 21) to give benzhydryl azoles 3t,u in only
50% and 5% yields, respectively. The substantially lower yield of 3u might reflect the
fact that 5-phenyl tetrazole 2d is far less active, although 2a and 2d reacted with xan-
thene 1a with comparable ease, probably owing to the leveling effect associated with the
remarkable reactivity of 1a (entries 1 and 4). Isochromane 1h (entry 22) and protected
1,2,3,4-tetrahydroisoquinoline 1i (entry 23), which possess a methylene moiety coacti-
vated by a phenyl group and an adjacent heteroatom, are less reactive than xanthene 1a
as well, and related coupling products 3v,w were produced in 70–74% yields. The use of
the methylene compound 1j bearing an aryl, a remote methoxy and an adjacent benzoy-
loxy group furnished in 30% yield bisazole 3x, which might arise from the further sub-
stitution by 2a of the initial C–N coupling product (entry 24). The extraordinary
performance of xanthenes and thioxanthenes as C(sp3)–H sources might be attributed
to the activation of the heteroatom-bridged bisaryl group, and the use of other diacti-
vated methylenes, such as 2-phenylacetophenone, ethyl benzoylacetate, 1,3-benzodioxole,
and 1,3-dithiane, failed to give desired C–N bond forming products (Fig. 1).