Journal of the American Chemical Society
Communication
Table 1. Optimization of Reaction Conditions
Entry
Variation from Standard Conditions
Yield (%)
a
1
2
3
4
5
6
7
8
none
1b in lieu of 1a
1c in lieu of 1a
58
a
49
n.d.
bc
,
HSiPh2(OtBu) in lieu of 1a
1a: 1.6 equiv in lieu of 1.0 equiv
1a: 2.0 equiv in lieu of 1.0 equiv
Quenching without warming up
37
87
b
b
b
94
<5
b
THF: 0.04 M in lieu of 0.02 M, otherwise under the
conditions of entry 6
TEMPO (3.0 equiv), otherwise under the conditions of
entry 8
98
93
97
b
b
9
10
11
9,10-Dihydroanthracene (3.0 equiv), otherwise under the
conditions of entry 8
iPrMgCl·LiCl (1.0 equiv) in lieu of BuLi, 1-I-2-(phenyl
n.d.
t
ethynyl)benzene in lieu of 2a, otherwise under the
conditions of entry 8
Et2O in lieu of THF
b
12
<5
a
b
Isolated yields. NMR yields determined using mesitylene as an
c
Figure 1. Toward nucleophilic activation of hydrosilanes: (a) our
strategy and (b) selectivity of ligand transfer in nucleophilic activation
of silanes, with ΔG values (kcal/mol) calculated at the level of
B3LYP/6-31+G*.
internal standard. The product has two phenyl groups on silicon as
exocyclic substituents. THF = tetrahydrofuran. n.d. = not detected.
the hydride (ΔG‡TS1 = 1.7 kcal/mol vs ΔGCP2‑CP4 + ΔG‡
=
TS2
17.7 kcal/mol) (Figure 1b). Thus, we adopted a strategy of
tying back two of the alkoxy ligands on the silicon by using
diols. Moreover, envisioning enthalpic activation of hydride,
we adopted a strain-imposing strategy with a multiply
methylated ligand for steric acceleration (Figure 1a-2). To
test this design strategy, we prepared seven- and six-membered
hydrosilanes containing diols equipped with multiple methyl
groups (1a−1c), as shown in Figure 1a-3. Notably, these
silanes, which were obtained in high yields, are stable to air and
moisture, in contrast to frequently employed diols such as
pinacol and 2,2-dimethyl-1,3-propanediol.14
Hydride transfer was very slow at −78 °C and warming to rt
was required for smooth progress (entry 7). The yield reached
the maximum under the conditions of entry 8. Use of 3 equiv
of TEMPO or 9,10-dihydroanthracene under the optimized
reaction conditions had no effect, and 3a was obtained in 93%
and 97% yields, respectively (entries 9 and 10). These results
strongly indicate that the present process does not involve free
radicals as key species promoting the reaction. Use of
magnesiated species only gave diphenylacetylene and un-
changed 1a, with no silicate formation (entry 11). Diethyl
ether basically shut down the reaction, whereas THF was
extremely effective (entry 12). Although tBuLi was found to be
the most suitable metalating agent in this system, we also
Next, we selected 2-phenylethynylphenyllithium, prepared
from bromoarene 2a, as a model substrate (Table 1). Upon
treatment of the aryllithium with 1a (1.0 equiv) hydride
transfer proceeded efficiently and subsequent cyclization gave
the desired benzosilole 3a in 58% yield (entry 1). The regio-
and stereoselectivity of hydrosilylation were perfectly con-
trolled, whereas such control is sometimes problematic in the
case of transition metal or Lewis acid catalysis. Use of 1b
slightly decreased the yield of benzosilole to 49% (entry 2). In
the case of 1c, the product was observed by GCMS analysis of
the crude mixture, but the silylether was cleaved during
chromatographic purification on silica gel (entry 3). The use of
HSiPh2(OtBu), which is involved in Lee’s hydrosilylation,8
afforded the corresponding benzosilole in 37% yield accom-
panied by the formation of the silylated diphenylacetylene in
n
s
evaluated less aggressive BuLi and BuLi. The use of these
alkyllithiums under optimized conditions afforded 3a in good
yield under noncryogenic conditions.
DFT calculations revealed that the activation energy for
hydride transfer from the silicon center to the carbon−carbon
triple bond is 15.0 kcal/mol with a large energy gain of −24.1
kcal/mol resulting from strain release (Scheme 2). The silicate-
forming step and the ring-closing step are highly reversible and
essentially barrierless processes with activation energy values of
1.8 and 2.4 kcal/mol, respectively (for the entire reaction
With the optimal set of reaction conditions in hand, the
scope of the benzosiloles was investigated (Table 2). This
reaction can be easily scaled up, and 3a was obtained on a
gram scale in 88% yield. The fluorescence quantum yield of 3a
t
12% yield via undesired substitution of BuO ligand (entry 4).
The poor ligand transfer selectivity observed here can be
attributed to the lack of entropic stabilization of the
pentacoordinate silicate with the monodentate alkoxide ligand
upon activation with the highly basic carbanion. A higher
loading of 1a improved the yield of 3a (entries 5 and 6).
4880
J. Am. Chem. Soc. 2021, 143, 4879−4885