.
Angewandte
Communications
PdCl2 and PtCl2, were also reduced to give metallic Pd and
Pt particles, respectively (entries 5 and 6), although a longer
reaction time was required. CrCl3(thf)3 reacted with 0.5 equiv
of 1 to generate green mixture containing CrCl2(thf)2
(entry 7), whereas MnCl2(thf)2 was inert due to its highly
negative redox potential (entry 8).
As compound 1 could be employed in the salt-free
reduction of some late-transition-metal chlorides to form
the corresponding metal particles, we next examined the
catalytic activities of the metal particles in Ullmann coupling
reactions of 4-iodoanisole (see the Supporting Information,
Table S3), which was recently applied to noble metal nano-
particle catalysts.[8,16] Upon combination with a slight excess
of 1 (1.5 equiv), NiCl2 and CoCl2(thf)1.5 in C6D6 at 808C
showed strong catalytic activities, giving 4,4’-dimethoxybi-
phenyl in 68% and 92% yields, respectively. In sharp
contrast, Pd0 and Pt0 particles generated in situ did not exhibit
catalytic activity in Ullmann coupling reactions (22% and 0%
from Ni(acac)2/1 in toluene at 808C, exhibited almost the
same catalytic activity in this reaction (entry 5). To our
surprise, no catalytic activity was observed for three different
sets of Ni nanoparticles generated from known procedures
using NiCl2/Li/4,4’-di-tert-butylbiphenyl, Ni(acac)2/NaBH4,
and Ni(cod)2/H2, all of which were typical experimental
conditions used for the preparation of crystalline Ni nano-
particles (Ni cNPs; entries 6–8).[18,19] Additionally, when
NaBH4 was employed as a reductant in place of 1 (Table 2,
entry 9), it was found that the reaction was unsuccessful
(using Ni NPs generated by Ni(acac)2/1), indicating that 1 was
a superior reductant to NaBH4 in this biaryl formation
reaction.
To gain insight into the catalytically active Ni species, we
characterized the Ni species generated in situ in the absence
of any substrate and ligand by transmission electron micros-
copy (Figure 1). The observed image clearly showed the
yields, respectively), although they were superior to nonpre-
2
À
cious catalysts for activating C(sp ) I bonds. The other metal
particles listed in Table 1 did not catalyze biaryl formation.
Further screening of Ni(acac)2 and Co(acac)2 led us to
determine that Ni(acac)2 and 1 was the best catalytic system
for the homocoupling reaction of 4-iodoanisole.
The optimized conditions established for the homocou-
pling of 4-iodoanisole using Ni catalysts were also applicable
to 4-bromoanisole, converting it into 4,4’-dimethoxybiphenyl
in 96% yield (Table 2, entry 1). The catalyst loading could be
decreased to 1 mol% and the reaction afforded the product in
94% yield (entry 2). Addition of TMP, a product that formed
from 1, showed no accelerating effect on the reaction to
improve the yield of the product (entries 3 versus 4). This
result indicated that TMP did not act as a ligand for the Ni
species, although N-heteroaromatic compounds, such as
pyridine and 2,2’-bipyridine, were reported to work as ligands
for some Ni-catalyzed reactions.[17] In fact, Ni nanoparticles,
which had been generated before the homocoupling reaction
Figure 1. a) High-magnification TEM image and b) corresponding
transmission electron diffraction (TED) pattern of Ni aNPs generated
from Ni(acac)2/1 (entry 5 in Table 2). c) High-magnification TEM
image and d) corresponding TED pattern of Ni cNPs generated from
NiCl2/Li/4,4’-di-tert-butylbiphenyl (entry 6 in Table 2). Scale bars in
TEM images (bottom right)=10 nm.
formation of Ni particles with an approximate size of 15 nm
(Figure 1a), and the corresponding electron diffraction pat-
tern appeared as a halo, suggesting that the nanoparticles
were amorphous in nature (aNPs; Figure 1b).[20] The crystal-
linity and size of these Ni aNPs remained intact after the
biaryl formation reaction (Figure S2). On the other hand,
a clear electron diffraction ring pattern was obtained for Ni
particles with an approximate size of 8 nm derived from
NiCl2/Li/4,4’-di-tert-butylbiphenyl, which were isolated as Ni
cNPs (Figure 1c and d). When NaBH4 was used as reductant,
Ni NPs approximately 9 nm in size with low crystallinity were
generated (Figure S2). These results provided the first
evidence that the catalytic activity of Ni NPs for reductive
biaryl formation depended on the crystallinity of the Ni NPs,
although it has been reported that Pd aNPs exhibit higher
activity for Suzuki–Miyaura coupling reactions than the
corresponding cNPs.[21]
Table 2: Catalytic performances of different Ni sources for the Ullmann
coupling reaction.
Entry
Ni source (mol%)
Yield (%)[a]
1
2
Ni(acac)2 (5)
Ni(acac)2 (1)
Ni(acac)2 (5)
Ni(acac)2 (5)
Ni particles from Ni(acac)2/1 (5)
Ni particles from NiCl2/Li/DTBB[d] (5)[f]
Ni particles from Ni(acac)2/NaBH4 (5)[f]
Ni particles from Ni(cod)2/H2(3 atm) (5)
Ni particles from Ni(acac)2/1 (5)
96
94
34
27
94
0
3[b]
4[b,c]
5
6
7
0
0
We further characterized the Ni aNPs by X-ray photo-
electron spectroscopy (XPS; Figure S3). In the XPS spectrum,
a major sharp peak for Ni 2p3/2 (852.5 eV) was detected (and
assigned to Ni0) with a relatively weak peak for Ni 2p3/2
(854.7 eV; attributed to NiO). In contrast, the XPS spectrum
of Ni cNPs generated from NiCl2/Li/DTBB displayed one
8
9[e]
trace
[a] Yield of isolated product. [b] Reaction time was 9 h. [c] TMP
(10 mol%) was added as additive. [d] DTBB=4,4’-di-tert-butylbiphenyl.
[e] NaBH4 (1.25 equiv) was used as reductant instead of 1. [f] In THF,
instead of toluene.
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 14437 –14441