Organic Letters
Letter
derivatives,23and other related EnT processes have recently
emerged.28 In their system, a bis-trifluoromethyl benzoyl
activating group was necessary to lower the triplet energy (ET)
of the precursor for effective energy transfer (EnT). In stark
contrast, while developing triplet nitrene C−H amination
reactions using analogous urea derived precursors, we had
identified that precursors containing a bis-trifluoromethyl
benzoyl activating group were completely ineffective to form
the rare class of carbamoyl nitrenes (Scheme 1B). Herein, we
report a photocatalytic strategy to form such triplet nitrenes
from engineered O-acyl hydroxylamine precursors that enables
new photocatalytic C−H amination reactions forming
imidazolidinones and related heterocycles (Scheme 1C). Our
preliminary results clearly highlight the subtle yet significant
differences that hydroxylamine nitrene precursors impart on
photocatalytic C−H amination reactions.
a
Table 1. Optimization of Acyloxy Urea C−H Amination
entry
sm
solvent
temp (°C)
base
catalyst yield (%)
1
2
3
4
5
6
7
8
9
10
1a
1b
1a
1a
1a
1a
1a
1a
1b
1c
1d
1d
1d
1d
(CH2Cl)2
(CH2Cl)2
CH2Cl2
CH2Cl2
(CH2Cl)2
(CH2Cl)2
MeCN
MeCN
MeCN
MeCN
MeCN
100
120
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
TMG
TMG
K2CO3
K2CO3
K2CO3
K2CO3
Et3N
Et3N
Et3N
Et3N
Et3N
Et3N
I
86
69
16
9
10
0
60
99
65
0
II
III
IV
V
b
b
VI
VI
VI
VI
VI
VI
b
Building on our work on heteroatom-substituted isocyanate
cascade reactions,29 we initially envisioned a reaction in which
thermal substitution on an acyloxy O-isocyanate followed by α-
elimination to a singlet nitrene could lead to nitrogen-rich
heterocycles (Scheme 2). Although some success was achieved
b
b
11
12
13
14
91
0
c
MeCN
MeCN
MeCN
b
0
0
b
Et3N
a
Scheme 2. Initial Thermal Cascade Reaction Design and
Associated Challenges
Conditions: Starting material (sm) 1a−d (1.0 equiv), base (1.1
b
equiv), catalyst (2 mol %), solvent (0.2 M), 16 h under argon. 450
c
nm blue LEDs, rt, 16 h. No light. 1H NMR yields (1,3,5-
trimethoxybenzene (TMB) as an internal standard).
transfer (SET)-initiated N-centered radical pathway.31 Under
identical conditions, we achieved 60% yield (entry 7). Upon
screening a variety of conditions, Ru(bpy)3(PF6)2 emerged as a
uniquely proficient photocatalyst in the presence of triethyl-
amine as a base (entry 8).
At this point, we investigated the scope of the reaction (e.g.,
entry 9, see Figure S1), which revealed systematic issues. The
most problematic byproduct was identified as II (Scheme 2)
formed from reduction of the N−O bond (ca., 10−20% yield).
This prompted a return to substrate design and a systematic
probe of the acyl substituent. Reasoning that SET reduction
could occur from RuI (E1/2 RuII*/RuI = −1.33 V) in the
presence of triethylamine (E1/2 = −0.78 V) as a sacrificial
reductant,32 a variety of O-acyl hydroxylamines were screened.
Notably, a substrate possessing the same bis-trifluoromethyl
benzoate successful in Chang’s work failed to provide any
desired product (reagent 1c, entry 10).23 In stark contrast, a
substrate bearing an electron-rich p-methoxybenzoate proved
effective (reagent 1d, entry 11). Control experiments verified
the need for light, base, and catalyst (entries 12−14).
The scope of the reaction was again investigated, and
contrary to results obtained with benzoate derivatives, using
electron-rich p-methoxybenzoates as substrates resulted in
efficient amination of 3°, 2°, and aromatic C−H bonds (Figure
1). Tertiary C−H bonds were aminated efficiently (2a−2o).
C−H amination at α-hydroxy/α-ethereal positions led to the
corresponding aromatic imidazolones (2l−2n), highlighting
the versatility of this method beyond the synthesis of saturated
heterocycles. A hydantoin heterocycle (2o) was also obtained
by a targeted C−H amination at an acetal methine C−H.
Unfortunately, with these conditions, the p-methoxybenzoate
precursor was still unable to effect C−H amination using
analogous amide and carbamate precursors (2g−2i) in
acceptable yields; however, modification of the base allowed
extension to these classes of nitrene precursors.14,23 Pleasingly,
efficient C−H aminations on unactivated methylenes, which
in developing a cascade process (see SI), several competing
side-reactions, such as O-isocyanate trimerization, pre-emptive
N−O bond cleavage, and 1,2-rearrangement followed by N-
isocyanate dimerization,30 obstructed the development of this
cascade (Scheme 2, I−III). However, some side-reactions were
mitigated by a stepwise approach (Table 1, entry 1).
Unfortunately, these conditions were not broadly applicable
(entry 2, see SI for preliminary reaction scope) but clearly
establish where the metal-free background reaction lies. To
minimize the decomposition pathways occurring from thermal
activation, extensive efforts using transition metal catalysts
were undertaken (entries 3−6, see SI for further expts).
Ultimately, catalysts reported to perform nitrenoid C−H
aminations typically catalyzed the 1,2-rearrangement side-
reaction, which was followed by rapid N-isocyanate dimeriza-
tion30 (see SI). We reasoned that selectively accessing the
triplet nitrene could provide a means to circumvent the 1,2-
rearrangement and enable productive C−H amination
reactions with free nitrenes. Toward this, we were drawn to
photoredox catalysis, intrigued by a report wherein acyloxy
ureas underwent intramolecular aziridination reactions, using
an Ir photoredox catalyst and a proposed single electron
B
Org. Lett. XXXX, XXX, XXX−XXX