Organic Letters
Letter
decarboxylative procedure12 was pioneered by Altman12a−c
with excellent regioselectivity in the presence of copper
catalysts (Scheme 1c). Consequently, the exploration of
metal-free and inexpensive photocatalytic procedures13 for
the dehydroxylative trifluoromethylation of allylic alcohols was
enthusiastically pursued. With our continuous interest in C−
OH bond cleavage and related green transformation,14 we
developed an organic photoredox-catalyzed dehydroxylative
trifluoromethylation of electron-withdrawing group activated
allylic alcohols (Scheme 1d), which would be complementary
work to previous investigations. The desired product equipped
with ester groups provided facile access to distinct molecules
that previous methods could not generate. In this reaction,
readily available CF3SO2Na15 was selected as the trifluor-
omethylation reagent. Under organic photoredox catalysis, in
situ generated byproduct SO2 was reutilized to activate the C−
OH bond, which enabled the reaction to occur through an
SN2′ process under mild conditions.
higher yield, while the isomer ratio was as low as 90/10 (E/Z).
Further investigation revealed that the reaction was sensitive to
the solvent. Only diminished desired products were obtained
when CH3CN was replaced by other solvents such as DMF,
THF, toluene, etc. (for details, see Supporting Information
Subsequently, the generalizability of this reaction was
evaluated using a variety of allylic alcohols (Scheme 2). Allylic
alcohols with a wide variety of substituents on phenyl ring were
found compatible with this transformation, delivering corre-
sponding allylic CF3 in good to high yields. Functional groups
such as halide, CF3, CN, NO2, and CHO were well tolerated.
Both electron-withdrawing (3b−3h) and electron-donating
(3i−3k) aryl substituted allylic alcohols were amenable to this
protocol. The positions of the substituents (at the para or meta
positions) on the phenyl ring have limited effects on the overall
transformation (3b−3s). In addition to monosubstituted
versions, allylic alcohols with multiple substituents on the
aromatic ring were compatible with the reaction (3t, 3u).
Fused-aromatic allylic alcohols could be utilized in this
reaction, generating 3v in a high yield. Moreover, allylic
alcohols bearing heteroaryl substituents, such as pyridien-2-yl
(3w) and thiophen-2-yl (3x) were also well tolerated well. In
addition, methyl (2E,4E)-5-phenyl-2-(2,2,2-trifluoroethyl)-
penta-2,4-dienoate (3aa) can be obtained by selecting
cinnamenyl α-substituted allylic alcohols, albeit the yield was
slightly lower. Besides aryl substituted allylic alcohols, alkyl
substituted versions also worked well under identity conditions
(3ab, 3ac). Double trifluoromethylation was performed well by
using ethyl 3-(4-(2-(ethoxycarbonyl)-1-hydroxyallyl)phenyl)-
2-hydroxybut-3-enoate (3ad). The structure of 3ad was
determined by X-ray analysis (3ad, CCDC 2071921).
Remarkably, γ-blocked allylic alcohols also could also
participate into this transformation, albeit affording 3ae with
a lower yield. Further investigation indicated that the electron-
withdrawing group on the β-position of allylic alcohol was
crucial to this transformation (3af−3ah). For instance, no
reaction occurred when 2-methyl-1-phenylprop-2-en-1-ol was
selected as the substance (3ag). Moreover, an alkynyl group on
the phenyl ring was compatible with the reaction conditions,
affording 3ai in 57% yield. This observation allowed for further
functionalization with a Cu-catalyzed click reaction. Moreover,
the reaction could also be performed at gram scale, with 3a
(72%, 1.76 g) and 3ai (44%, 1.01 g) isolated in comparable
yield. Without other notice (3s, 3x, and 3af), the final product
was detected with excellent E-selectivity. Unfortunately, this
catalytic system was found inefficient to primary allylic
alcohols, such as methyl (E)-2-(hydroxymethyl)-3-phenyl
acrylate or (E)-2-nitro-3-phenylprop-2-en-1-ol (for details,
In an attempt to access the desired allyl-CF3 compounds
(3a), allylic alcohol (1a) and CF3SO2Na (2a) were selected as
the model substrates. At room temperature, a variety of
photoredox catalysts were screened in CH3CN solvent (Table
1). After 24 h, only a trace amount of 3a or even no 3a was
a
Table 1. Screening Reaction Conditions
e
b
c
entry
cat. (mol %)
eosin Y (1)
t (h)
3a (%)
E/Z
1
2
3
4
5
6
7
8
9
24
24
24
24
24
24
10
4
0
f
f
32
26
68
67
73
77
0
Ru(bpy3)Cl2 (1)
Ru(bpy3)(PF6)2 (1)
PC-1 (1)
95/5
>99/1
>99/1
>99/1
>99/1
90/10
PC-2 (1)
Mes-Acr+Ph(BF4−) (1)
Mes-Acr+Ph(BF4−) (2)
Mes-Acr+Ph(BF4 ) (4)
4-CzIPN (1)
Mes-Acr+Ph(BF4−) (4)
‑
1.5
4
d
10
a
Experimental conditions: 1 (0.3 mmol), 2a (0.45 mmol), and cat
b
were mixed in CH3CN (4.5 mL) under blue-LED (18w*3). Isolated
yield. Determined by crude H NMR. Without light sources.
c
d
e
1
The reproducibility of this protocol was also evaluated by
using the methodologies reported by Glorius et al.18 Factors
such as concentration, oxygen level, scales, water level,
temperature, and light intensity was screened and compared
with the standard condition (for details, see SI, Sensitivity
assessment). Among these parameters, except for the higher
temperature having limited effect on this transformation, all the
other variations caused only negligible changes. Therefore, this
investigation indicated this reaction has good reproducibility.
The synthetic utility of this transformation was then
preliminarily studied. We initially examined late-stage trifluor-
omethylation toward several allylic alcohols bearing bio-
logically active skeletons. As is shown in Scheme 3, L-
detected when Ru(bpy3)Cl2, Ru(bpy3)(PF6)2, and Eosin Y
were selected as the catalysts (Table 1, entries 1−3). To our
delight, acridinium ion photoredox catalysts16 PC-1 (Table 1,
entry 4) and PC-2 (Table 1, entry 5) delivered 3a in
moderated yields with high E/Z ratios. Mes-Acr+Ph(BF4 )
−
gave a more positive result in terms of higher yield and ratio of
isomers E/Z (Table 1, entry 6). Increasing the catalyst loading
(Table 1, entries 7 and 8) enabled the transformation to be
finished in 4 h with 3a isolated in 73% yield (Table 1, entry 8).
In the presence of catalyst 4-CzIPN,13a,b,17 3a was obtained in
5236
Org. Lett. 2021, 23, 5235−5240
Li, Bowen
