Organic Letters
Letter
Langlois reagents.22−26 Well-designed sources of CF3/CF2H are
still not available, as recent studies have highlighted the
problems with the currently available ones.27,28 Moreover, we
previously reported a useful method for carrying out
trifluoromethylations/cyclizations of 1,7-enynes by using
Ph2SCF3OTf as a CF3 source in continuous flow.29 In 2017,
Akita and co-workers reported a shelf-stable easy-to-handle
sulfonium salt, namely, Ph2SCF2HBF4 (S-difluoromethyl-S-
di(p-xylyl)sulfonium tetrafluoroborate).30
a
Table 1. Optimization of the Reaction Conditions
b
entry
deviation from standard conditions
yield (%)
1
2
3
none
83
17
53
28
21
40
51
0
The use of light energy impairs the overall sustainability of the
photocatalytic reactions as dictated by the Bouguer−Lambert−
Beer law. Photochemical processes and continuous-flow
chemistry have, when used in combination, led to the
development of various reactions.31−33 In continuous-flow
chemistry, continuous reactions are carried out in tubes or a
microreactor. The keys to the success of carrying out reactions
using continuous flow have been shown to involve achieving an
improved mixing ability and more efficient heat transfer as well
as the ability to scale up such reactions.34 Compared to the
traditional batch reactions, continuous flow technology offers
many advantages. This technology has been used in the
screening and optimization of various reactions.35,36 Jamison
and Stephenson et al. have realized their unprecedent works in
this area.37 However, to the best of our knowledge, there has
been no report on the construction of indole[2,1-a]isoquinoline
derivatives in continuous flow under the condition of metal-free
photocatalysts and green solvent without any other bases.
Inspired by these achievements, we set out to and successfully
developed photocatalyzed tri- and difluoromethylation/cycliza-
tion process for the construction of a series of tri- or
difluoromethylated indole[2,1-a]isoquinoline derivatives in
good yields under mild conditions via a noble metal-free
protocol in continuous flow.
Initially, we commenced our study by employing (S)-3,5,12-
trimethyl-5-(2,2,2-trifluoroethyl)indolo[2,1-a]isoquinolin-
6(5H)-one (1a) and diphenyl(trifluoromethyl)sulfonium tri-
fluoromethane-sulfonate (2a) as the model set of substrates in
CH3CN to test the reaction conditions (Table S1, entry 8).
Although the desired product 3a was observed, the reaction
generally proceeded slowly. Thus, we set up a continuous-flow
photoreactor (Figure S2) made according to the needs of the
reaction to speed up the efficiency of the transformation (Table
S1, entry 7). Next, we investigated solvents of the reaction. To
find lower toxicity and less costly green solvent, a series of other
solvents including THF, DMSO, MeOH, 1,4-dioxane, and DMF
were screened, but acetone was indicated from the results to be
the best choice for this reaction (Table S1, entries 1−12). The
use of EtOH reduced the onset potential of reaction due to its
poor ability to dissolve both photocatalyst and reagents (Table
1, entry 3). To our delight, the desired indole[2,1-a]isoquinoline
(3a) product was obtained in 10 min in 83% yield in acetone at
room temperature when using the organic photocatalyst (4s,6s)-
2,4,5,6-tetra(9H-carbazol-9-yl)isophthalonitrile (4CzIPN)13 as
the photocatalyst, a PFA tube (ID = 600 μm, volume = 1.0 mL),
a flow rate of 100 μL/min, and 50 W blue LED irradiation
(Table 1, entry 1). In further optimization studies, fac-Ir(ppy)3
displayed significantly lower catalytic efficacy (Table 1, entry 7).
However, when Ph2SCF3OTf was replaced with CF3SO2Na as
the source of CF3, the yield of the product 3a decreased to 28%
while extra water was needed to achieve a clear reaction solution
(Table 1, entry 4); the poorer yield was attributed to the
presence of this extra water. The addition of bases resulted in less
reactivity for this reaction (Table 1, entries 6 and 7). From
MeOH as the solvent
EtOH as the solvent
CF3SO2Na as the CF3 source
Et3N as additive (2.0 equiv)
pyridine as additive (2.0 equiv)
fac-Ir(ppy)3 as the catalyst
in batch
c
4
5
6
7
d
8
9
without catalyst
2
10
11
12
two injection system
tube (ID = 800 μm)
200 μL/min
76
75
49
a
Standard reaction conditions: 1a (0.2 mmol), 2a (0.22 mmol),
4CzIPN (1.0 mol %) in 2.0 mL of acetone irradiated with light from
50 W blue LEDs (455 nm) at room temperature inside a 1.0 mL
reactor of a flow system (PFA tube, ID = 600 μm, 100 μL/min). See
each recrystallized product was calculated based on 1a. CF3SO2Na as
the CF3 source; here another 0.5 mL of H2O was needed for
achieving a dissolution of the reaction mixture. Air, 5 h.
b
c
d
control experiments, both light and the photocatalyst were
concluded to be crucial to the success of the reaction (Table S4,
entries 1−4). Using a batch condition, however, failed to deliver
product 3a by 19F NMR, even when a longer reaction time was
used (Table 1, entry 8). We speculate the continuous flow
system was the most efficient as well as mixing ability compared
to the batch. However, the mechanism of the phenomenon is
not clear. The flow conditions were also examined, and use of a
thicker pipe (ID = 800 μm) and faster rate (200 μL/min)
provided lower yields (Table 1, entries 11 and 12).
With optimized conditions in hand, we investigated the
scopes of indole substrates and CF3 sources for the photo-
catalytic di/trifluoromethylation/cyclization reaction. As shown
in Scheme 2, a variety of functional groups were observed to be
well tolerated, forming products bearing various substituted 2-
arylindoles in the flow system. Notably, changing the electronic
properties of the substituent (R1) at the para position of the 2-
phenyl moiety of the N-substituted 2-aryl indole was found to
have little influence on its catalytic efficiency. Electron-neutral
substituents (−Me), electron-withdrawing groups (−F, −Cl,
−Br), and an electron-donating group (−OMe) were all
compatible with this reaction, giving the corresponding products
in 52−83% yields (3a−3g). Subsequently, the electronic nature
of the substituent located on the C3 position of the indole ring
was investigated; for two cases, products formed efficiently
regardless of the electronic nature of the substituent (3h and 3i).
Various substituents at the C5 position of the indole ring were
also readily tolerated in this system, delivering the correspond-
ing products in good to moderate yields (3j−3n). Besides these
examples, substrates containing different groups at the R4
position also gave the corresponding products in moderate
yields (3o−3q).
Due to their special chemical and biological properties,
difluoromethyl compounds are significant and widely used in the
1951
Org. Lett. 2021, 23, 1950−1954