Organic Letters
Letter
a
Scheme 2. Synthesis of 4-Fluoroacridines via Cascade
Radical Cyclization
Table 1. Optimization of Reaction Conditions
3
4
b
b
entry
photocatalyst
base (equiv)
none
solvent
(%)
(%)
c
1
none
Ru(bpy)3Cl2
MeCN
MeCN
NR
37
39
55
c
2
nBu3N (0.2)/K2CO3
(2)
3
Ru(bpy)3Cl2
nBu3N (0.2)/K2CO3
(2)
MeCN
49
94
4
Ru(bpy)3Cl2
Ru(bpy)3Cl2
Ru(bpy)3Cl2
Ru(bpy)3Cl2
nBu3N (3)
K2CO3 (3)
Et3N (3)
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
DMSO
DMF
58
NR
65
71
67
56
55
60
57
43
83
74
11
69
83
5
6
7
122
80
76
iPr2NEt (3)
redox catalysis. We envisioned that the use of vinyldiazo
reagents to replace alkenes as the radical acceptors would lead
to 4-fluoroacridines. The working mechanism is shown in
Scheme 2. The addition of difluoromethyl radical A to
vinyldiazo compounds generates vinyl radical B, which
undergoes addition to C−C triplet bond of ketoimine 1
intramolecularly to give vinyl radical C. The intramolecular
homolytic aromatic substitution (HAS) of radical C, followed
by single electron oxidation and elimination of HF would
produce 4-fluoroacridine 3. Due to the higher bond
dissociation energy (BDE) of C−Cl bond, the termination of
vinyl radical B by a halogen atom transfer can be suppressed.
However enticing this proposal might appear, it was conceived
to have two major challenges: (1) How will initial radical A be
generated if less reactive C−Cl was used? (2) If photocatalyst
was employed to generate radical A, how will the background
pyrazole formation mediated by light or base be suppressed?13
Here, we report the results of our research.
8
Ru(phen)3Cl2 iPr2NEt (3)
83
9
Ir(ppy)3
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
iPr2NEt (3)
89
10
11
12
13
14
15
16
Ir(dFppy)3
Eosin Y
92
83
Ru(bpy)3Cl2
Ru(bpy)3Cl2
Ru(bpy)3Cl2
Ru(bpy)3Cl2
Ru(bpy)3Cl2
Ru(bpy)3Cl2
96
106
70
76
DCM
DCE
toluene
DCM
119
41
c,d
17
a
Reaction conditions: 1a (0.2 mmol), 2a (0.5 mmol), base,
photocatalyst, and solvent (1.5 mL) under the irradiation of a 5 W
b
blue LED at room temperature for 18 h. Yields of 3a and 4 were
determined by 1H NMR using 1,3,5-trimethoxybenzene as the
c
d
internal standard. 1.5 equiv of 2a. 5 mmol scale reaction, and 2a
was added in three portions.
In our previous report on radical 1,3-addition of
perfluoroalkyl iodides to vinyldiazoacetates, the initial Rf
radical was generated by forming a radical pair complex
between RfI and triplet free vinylcarbene.5 However, the direct
generation of gem-difluoromethyl A in the absence of
photocatalyst is less like due to the high dissociation energy
of C−Cl bond. As expected, when the MeCN solution of
alkynyl ketoimine 1a and ethyl vinyldiazoacetate 2a was
irradiated with a 5 W blue LED at room temperature, no
desired 4-fluoroacridine 3a was detected, while the conversion
of 2a to pyrazole was found as the major background reaction
(Table 1, entry 1). Therefore, we investigated the reaction
whereas more polar solvents such as DMSO and DMF and less
polar toluene were found to be unfavorable (Table 1, entries
12−16). In a gram scale reaction, adding 1.5 equiv of 2a in 3
portions over the course of the reaction produces 3a in the
yield of 69%, and byproduct pyrazole 4 could be minimized to
41% yield (Table 1, entry 17).
With the optimized conditions in hand (Table 1, entry 14),
the generality of this cascade radical cyclization reaction for the
synthesis of 4-fluoroacridine was explored using a variety of
alkynyl ketoimines 1b−1s and ethyl vinyldiazoacetate 2a
(Scheme 3). Initially, the effects of the substitutents at the
para-position of phenyl ring connected to the C−C triple bond
were evaluated. Substrates bearing methyl, halogen, and phenyl
groups worked well, affording the desired products 3a−3f in
71−79% yields. The reaction became sluggish when strong
electron-withdrawing nitro and electron-donating methoxy
groups were introduced (3g and 3h). The presence of a methyl
group at the meta-position and a fluoro group at the ortho-
position had no apparent effect on the reaction. When R1 was a
1-naphthyl and 3-thienyl group, the corresponding 4-
fluoroacridines 3k and 3l were obtained in the yields of 73
and 78%, respectively. It is worth mentioning that 9-
cyclopropyl-4-fluoroacridine 3m could be prepared by this
radical transformation. In addition, a terminal alkyne motif also
survived (3n), and no side product from radical addition to the
terminal CC bond was detected. Alkynyl ketoimines bearing
a substituent on the phenyl ring attached to the nitrogen atom
all reacted efficiently with 2a to give 7-bromoacridine 3o, 7-
n
using Ru(bpy)3Cl2 as the photocatalyst and Bu3N/K2CO3 as
the bases, which are the optimal conditions of our previous
report for the radical cyclization of 1a with alkenes. Although
the addition of photocatalyst and base favored the light-
mediated pyrazole formation, we observed the generation of 3a
in 37% yield (Table 1, entry 2). Increasing the amounts of 2a
to 2.5 equiv led to 3a in 49% yield, despite the yield of pyrazole
n
4 increasing to 94% (Table 1, entry 3). The use of Bu3N as
the sole base afforded 3a in 58% yield, while the replacement
n
of Bu3N by inorganic base K2CO3 shut down the reaction
i
completely (Table 1, entries 4 and 5). NEt3 and Pr2NEt gave
better yields compared with nBu3N (Table 1, entries 6 and 7).
Several photocatalysts, such as Ru(phen)3Cl2, Ir(ppy)3,
Ir(dfppy)3, and Eosin Y, were examined, but they were all
less efficient than Ru(bpy)3Cl2 (Table 1, entries 8−11).
Further inspection of the reaction conditions revealed that
dichloromethane (DCM) was the best solvent for the reaction,
4280
Org. Lett. 2021, 23, 4279−4283