Organic Letters
Letter
−
Scheme 2. Outline of Aminomethylation/Ring Opening of
vs SCE, E1/2(PS*/PS ) = +1.35 V vs SCE) and its lower cost
17
compared with iridium-based photocatalysts. In the presence
of 2 mol % 4-CzIPN, 5 mol % [Ni(bpy)Cl ], 10 mol %
Zn(OTf) , and DBU (0.5 equiv) in DMF (0.1 M) at room
2
2
temperature under blue-light irradiation for 24 h, 2a reacted
with 1a to furnish the alkylated ring-opening adduct 3a in 90%
conversion and 80% isolated yield (entry 1, Table 1). The
reductive ring-opening product 3′a was also obtained under
these conditions in 10% conversion. The side-product
formation might be explained by a competitive reaction
inevitable prefunctionalization of substrates and the formation
of the undesirable toxic wastes, and (ii) it would open a
straightforward access to aminomethyl dihydronaphthalenols,
whose synthesis is still a challenge in organic chemistry.
At the outset, we chose the aminomethylation of
oxabenzonorbornadiene 1a (1 equiv) with dimethylamine 2a
conditions (Tables 1 and S1−S6). On the basis of previous
+
between a nickel(0) complex and an acid (DBU·H ) to
generate a nickel hydride species. Without light, photocatalyst,
or nickel catalyst, no reaction occurred (entries 2−4, Table 1).
These results ruled out the free radical addition, as suggested
by our initial observations. The ligand on the nickel played a
crucial role in the catalytic activity. Replacing the bipyridine
(bpy) ligand by a phosphine led to a decrease in both the
reactivity and the selectivity (entries 5 and 6, Table 1). The
reductive ring-opening adduct was favored with phosphine
ligands (entries 5 and 6, Table 1). The electron-deficient
1
0
reports, 4-CzIPN was selected as the photosensitizer (PS)
because of its ability to oxidize alkylamines (E1/2 = +1.1−1.2 V
a
Table 1. Optimization of the Reaction Conditions
dipyridine amine dCF bpy also provided the alkylated ring-
3
opening product in lower yield, while the often-used dtbbpy
ligand furnished both ring-opening adducts 3a and 3′a with
lower conversion (85%) and lower selectivity (entries 7 and 8,
Table 1). In addition to nickel complexes, palladium and
cobalt precatalysts have also been introduced in dual
14
catalysis. Although Rovis recently reported the amino-
methylation of dienes via photoredox/low-valent cobalt
1
8
catalysis, no ring-opening adduct was formed with a
palladium or cobalt complex as the cocatalyst (entries 10
and 11, Table 1). The amount of aniline 2a drives the
formation of 3a over 3′a to some extent, and an increase in the
amount of 2a favored the aminomethylated product (entries 1,
1
3, and 14, Table 1). Any other modification of the reaction
conditions (use of a ruthenium complex or 4-CzTPN as the
PS, of other solvents, bases, Lewis acids, etc.) had a negative
effect on both the conversion and the selectivity for 3a. A NOE
analysis confirmed the structure of 3a, including the regio- and
cis stereoselectivity.
With these optimized reaction conditions in hand, we
explored the scope of this unprecedented ring-opening
reaction. Various substituted oxabenzonorbornadienes 1 were
initially evaluated (Scheme 3).
Oxabenzonorbornadienes bearing electron-donating sub-
stituents (e.g., Me and MeO) within the aromatic fragment
provided the corresponding ring-opening adducts 3b−d in
good yields (63−68%), while electron-withdrawing substitu-
ents lowered the chemical yield (3e) (Scheme 3). Dibromo-
substituted oxabenzonorbornadienes were nonreactive under
these dual catalysis conditions (Scheme 3). In all of these
examples, apart from the formation of the reductive ring-
opening adduct, no dehydration product was observed, and the
cis-1,2-addition compounds were exclusively formed. The syn-
1
,2-adduct 3f was also obtained from the disubstituted
bridgehead oxabenzonorbornadiene, albeit in a modest 20%
yield and 44% conversion (Scheme 3). Unsymmetrically
substituted oxabenzonorbornadienes were also introduced.
The ring-opening adducts 3g−j were isolated in moderate to
good yields (50−70%) but mainly as 1:1 mixtures of the two
possible regioisomers, except for 3i, for which a 9:1 ratio was
observed (Scheme 3). The low control of the regioselectivity
might be explained by inefficient steric interactions between
the substituents and the nickel complex (see the proposed
a
General conditions: 1a (0.2 mmol), 4-CzIPN (2 mol %),
[
Ni(bpy)Cl ] (5 mol %), DBU (0.1 mmol), 2a (0.6 mmol), DMF
2
b
(
4 mL, 0.1 M), 34 W Kessil blue LED lamp, rt, 24 h. Conversions
1
c
were determined by H NMR analysis of the crude mixtures. The
yield in parentheses is based on the isolated product.
B
Org. Lett. XXXX, XXX, XXX−XXX
Diallo, Abdoul G.
