Organic Letters
Letter
ineffective, affording only C2-alkenylated byproduct 3a′
(Scheme 2, L1−L3). Then secondary phosphine oxide
Ph2P(O)H was tested, but only a trace amount of product
3a was detected (L4). Considering the vital influence of the
steric and electronic property of ligand substituents on the
catalytic activity, diverse phosphine oxides were optimized.
Pleasingly, when Cy2P(O)H L5 was used, desired product 3a
was produced in a yield of 14%. Albeit in low yield, this result
demonstrated the successful Ni-catalysis for the target
annulation. By switching the backbone from a secondary
phosphine oxide to a diamino-phosphine oxide, L6 led to an
increased yield of 30%. Then substituent screening revealed
that di-tert-butylethyldiamino phosphine oxide L8 offered a
higher yield of 52%, but aromatic substituted ligands provided
no annulation product at all (L9−L10). Notably, cyclohexanedi-
amine-derived ligands exhibited better performance (L11−L15)
and a yield of 66% was obtained (L13). Then, increasing the
loading of alkyne and AlMe3 afforded a 93% yield, and in this
case only a trace amount of byproduct 3a′ was detected. It is
worth noting that, with low-valent Ni species as the catalyst,
trimerization of alkyne was observed as well, and the amount
of benzene derivative byproduct varied with reaction
conditions, with the yield being 9% under optimized
conditions. Control experiments showed that removal of any
component from the combination of Ni, Al, and L13 led to no
product, suggesting the critical role of the PO-bound Ni−Al
bimetallic catalyst.
Scheme 5. Mechanistic Experiments and a Plausible
Mechanism
With the above-mentioned optimal reaction conditions
established, we turned to investigate the substrate scope with
respect to imidazole substrates (Scheme 3). Imidazoles with
electron-rich or electron-poor substituents on the N-aromatic
ring all were competent substrates, offering aza-quinolines in
58%−95% yields (3b−3n). Functional groups including Me,
t
Et, Bu, OMe, OCF3, Ph, and CF3 were all well tolerated.
Scheme 6. Optical Property of the aza-Quinolines: (a)
Fluoresence Emission Spectra in CH2Cl2 Solution; (b)
Fluorescence Images of aza-Quinolines in Both the Solid
State and CH2Cl2 Solution under UV Light
Notably, satisfactory yields were retained in the presence of
ester group CO2Et, demonstrating the elegant recognition
ability of Al to the N atom of imidazole (3k, 3l). When the C3-
substituted N-aromatic ring was employed, C6−H activation
occurred due to steric hindrance (3c, 3d, 3h, 3l, 3n). And
substrates with C3-electron-donating groups generally pro-
vided higher yields, which suggested an electrophilic C−H
activation mechanism (3c, 3d, 3h). Besides, the pattern of
benzimidazole proved to be versatile as well. Benzimidazoles
decorated with different substituents all participated in the
annulation smoothly (3o−3u). Higher yields were observed
for electron-rich substrates compared with electron-poor ones,
likely due to the enhanced coordination ability to Al, which in
turn promoted the C−H activation of the imidazole ring (3p
vs 3u). Yields decreased with C4 or C7 substituted
benzimidazoles, probably because increased steric hindrance
impeded the second C−H activation step (3o, 3t).
Next, the compatibility of alkynes with this annulation
reaction was examined. As shown in Scheme 4, reactions took
place readily with a set of alkynes. Excellent yields ranging from
83% to 91% were observed for alkynes containing different
alkyl groups such as Et, nPr, and nhexyl (4a−4e). When
nonsymmetric alkyne was subjected to the standard conditions,
nearly 1:1 mixtures were furnished (4d, 4e). Aromatic alkynes
bearing H, Me, or Et were competent partners as well,
delivering desired products smoothly though in moderate
yields (4f−4i). Trimethyl(prop-1-yn-1-yl)silane was also tested
as protected terminal alkyne but gave no desired product.
4036
Org. Lett. 2021, 23, 4034−4039