Journal of the American Chemical Society
Communication
have never been reported so far,12,13 is urgent and in high
demand.
We hypothesized that the relatively strong London
dispersion interaction of the sterically hindered substituents
R on benzynes 6 at the 3-position should enable direct
regioselective (2 + 2) cyclodimerization reactions for accessing
proximal-7 in good yields (Scheme 2).
In the conversion of 5i, a significant amount (29%) of the
thia-Fries rearrangement side-product 8a was observed,14
which cannot be formed via benzyne 6i (Scheme 3). The 8a
Scheme 3. Inhibition of Thia-Fries Rearrangement
First, we generated a wide variety of 3-substituted benzynes
6 from 2-iodophenyl triflates 5 in the absence of arynophiles
using MeLi in THF at −78 °C (Table 1). Unsubstituted
Table 1. (2 + 2) Cyclodimerization Regioselectivities of 3-
a
Substituted Benzynes 6
formation indicated that the yield of proximal-7i could be
improved further by improving the benzyne generation from
precursor 5. After several trials, a new 3-adamantylbenzyne
precursor 5i′, bearing the p-toluenesulfonyloxy moiety, was
found as the leaving group and improved the yield of proximal-
7i to 74%, while maintaining complete proximal selectivity,
with the absence of the thia-Fries rearrangement product 8b
(Scheme 3). These varied results are likely due to the higher
bond strength of the O−S single bond of the p-
toluenesulfonyloxy group than that of the trifluoromethane-
sulfonyloxy group.
With these exciting results in hand, we hypothesized that
more attractive helical biphenylenes,15 such as proximal-10,
could be synthesized through the proximal-selective (2 + 2)
cyclodimerization of the fused tetracyclic benzyne 9. We
expected that the π−π interaction between two molecules of 9,
which is the key contributor of the dispersion force, would
work well as in the case of 6i (Scheme 4).16
b
ratio
yield
c
entry
R1
R2
5, 6
proximal:distal
7
(%)
d
1
2
3
4
5
6
7
8
9
H
H
H
5a, 6a
5b, 6b
5c, 6c
5d, 6d
5e, 6e
5f, 6f
5g, 6g
5h, 6h
5i, 6i
−
−
7a
7b
7c
7d
7e
7f
7g
7h
7i
20
0
e
OMe
Me
Me
Me
Me
Me
i-Pr
Me
Me
1.1:1
1.1:1
7.0:1
9.5:1
7.7:1
16:1
36
85
80
56
51
52
54
d
d
d
d
SiMe3
Si(t-Bu)Me2
Si(i-Pr)3
i-Pr
t-Bu
Ad
f
g
>20:1
a
Conditions: 1.0 equiv of 5, 1.5 equiv of MeLi in THF (0.10 M) at
b
−78 °C for 15 min. Ratio between proximal-7 and distal-7 was
c
determined by 1H NMR analysis of crude products. Combined
d
e
isolated yield of 7. Combined 1H NMR yield of 7. Complex mixture
f
g
was obtained. Ad = adamantyl. distal-7i could not be detected.
Scheme 4. Proximal-Selective (2 + 2) Cyclodimerization of
Benzo[c]phenanthryne 9 and Synthesis of Helical
Biphenylenes, proximal-10
benzyne 6a underwent (2 + 2) cyclodimerization to provide
biphenylene 7a in 20% yield (entry 1); however, 3-
methoxybenzyne 6b, which is known to give coupling products
in good yield with varieous arynophiles, did not afford
biphenylenes 7b (entry 2).12 The reaction of 3-methylbenzyne
6c gave proximal-7c and distal-7c in a 1.1:1 ratio in a combined
36% yield (entry 3). In contrast, a similar reaction with 3-
(trimethylsilyl)benzyne 6d produced the corresponding
biphenylene 7d in 85% yield, albeit with poor regioselectivity
(1.1:1) (entry 4). Eventually, we were delighted to see the
expected proximal selectivity (7.0:1) and good yield (80%)
when a tert-butyldimethylsilyl (TBDMS) group was adopted as
the directing group (entry 5). The reaction with the
triisopropylsilyl (TIPS) directing group exhibited in higher
regioselectivity (9.5:1) (entry 6).
Next, we surveyed the alkyl directing groups (entries 7−9).
The ratio of the proximal and distal biphenylenes 7g was 7.7:1
using 3-isopropylbenzyne 6g, and the combined yield of these
two products was 51% (entry 7). Notably, the proximal
selectivities improved according to the bulkiness of the R1 alkyl
substituent (entries 3, 7, and 8). Finally, the (2 + 2)
cyclodimerization of 3-adamantylbenzyne 6i provided prox-
imal-7i exclusively in 54% yield (entry 9). These results (Table
1) indicate that the London dispersion interaction between the
C3-directing groups successfully controls not only the
orientation of the reactions but also the product yields.
Before experimental verification of this hypothesis (Scheme
4), we conducted several theoretical studies (Figures 1 and 2).
The optimized structure of the desired product proximal-10a
(R1−R3 = H) and undesired distal-syn-10a and distal-anti-10a
were calculated at the B3LYP-D3(BJ)/PCM/def2-TZVP level,
where the Conductor-like Polarizable Continuum Model
(PCM) was adopted for including the solvent effects of
tetrahydrofuran (THF). The free energy difference among
them was evaluated at the DLPNO-CCSD(T)/def2-TZVP
level. Notably, the energy of the most sterically demanding
proximal-10a was the lowest and was 6.0 kcal/mol lower than
that of distal-anti-10a [Figure 1(a)]. While this large
stabilization of proximal-10a was unexpected, it would be
desirable for achieving regioselective (2 + 2) cyclodimeriza-
tions of fused benzynes 9. To rationalize the stability of
10854
J. Am. Chem. Soc. 2021, 143, 10853−10859