Organic Letters
Letter
a
radical combines with aromatic systems; finally, the resultant
radical undergoes oxidation to form the product. In principle,
the same chemical transformations could be accomplished
electrochemically by performing the redox events at the
cathode and anode sequentially. Because the radical
intermediates formed by CF3 radical addition to (hetreo)-
arenes lost the aromaticity, they are unstable30,31 and can cause
low yields of trifluoromethylated products using the paired
electrolysis setup, which we hypothesize will be addressed by
ACE.
Table 1. Reaction Development
b
c
conversion
(%)
yield
5/6
d
entry
Vp (V)
f (Hz)
(%)
ratio
1
2
3
4
5
6
7
8
9
0
n/a
100
100
100
100
100
10
1000
60
100
n/a
<1
6
3.3 (AC)
3.6 (AC)
4.0 (AC)
4.4 (AC)
4.8 (AC)
4.4 (AC)
4.4 (AC)
4.4 (AC, sine)
27
48
100
44
21
<1
40
66
13
Figure 2a shows our experimental setup. The power supply
was a waveform generator, which can supply either a constant
44
84
24:1
19:1
44:1
0.75:1
e
10
11
4.4 (AC)
4.4 (DC)
64
a
Reaction scale: 0.5 mmol of 2 (1 equiv) in 4 mL of acetonitrile.
b
c
Conversion and ratio of 5/6 were determined by 19F NMR. Isolated
d
e
yield of 5. Chlorinated product was isolated in 16% yield. 1 mmol of
2.
desired product 5 was obtained in a comparable yield (64%,
entry 10). To compare, the paired electrolysis condition only
gave a 13% yield of 5 using identical chemical reagents at a
constant voltage of 4.4 (entry 11). The control experiment
established that, when no voltage was applied, there was no
reaction (entry 1). Overall, the results in Table 1 confirmed
our hypothesis that the ACE significantly improved the yield
compared to the paired electrolysis.
Figure 2. (a) Photographs of the experimental setup for ACE. A
square waveform was applied between two glassy carbon plate
electrodes by a waveform generator. (b) Proposed mechanism for
electrochemical trifluoromethylation of 2-acetylpyrrole (2) using
ACE.
Next, we investigated the critical role of Vp and f in the ACE
method. We measured the standard reduction potential (E1) of
1 and the oxidation potential (E2) of 3, which determines the
voltage required for trifluoromethylation to proceed. The
cyclic voltammogram of 1 (Figure 3a) showed an onset
potential for 1 reduction of approximately −0.7 V vs Ag/AgCl.
We estimated E1 to be −0.97 V using the inflection points of
the cathodic wave.32 To detect the oxidation of the unstable
intermediate 3, we used fast-scan linear sweep voltammetry.33
We first held the electrode potential at −1.2 V to reduce 1 and
generate a pool of radical 3. Subsequent scanning the electrode
potential positively at a high scan rate of 20 V/s enabled the
oxidation of 3 before it proceeded through other pathways. At
low scan rates, the electrochemical signal from the oxidation of
3 significantly diminished (Figure S2), confirming the
instability of 3. We observed an anodic wave starting from 0
V with a major peak at ∼0.7 V (Figure 3b). When the potential
was held at −0.4 V, wherein no CF3 radical was generated, no
anodic wave was observed, suggesting the anodic wave results
from the oxidation of 3. To further confirm this result, we
carried out two additional experiments. First, we elongated the
holding time at −1.2 V to generate more CF3 radicals. Second,
we increased the concentration of 2 in the reaction mixture. In
both cases, we observed a significantly increased current
(Figure 3), confirming the anodic wave arose from the
oxidation of 3. Accordingly, we estimated E2 to be 0.5 V using
the inflection points of the anodic wave.32 Based on these
results, we estimated the thermodynamic voltage for electro-
chemical trifluoromethylation, |E1 − E2|, is ∼1.5 V. It is worth
noting that the current took off after ∼0.85 V for all
experiments due to direct electro-oxidation of 2.34−36
voltage bias (DC) to perform paired electrolysis or an
oscillating square waveform (AC) to perform ACE. During
ACE, the reaction was initiated by reducing 1 to CF3 radicals
when the potential of an electrode was negative (Figure 2b).
The generation of CF3 radicals was confirmed by a radical
trapping experiment using TEMPO (Figure S1). CF3 radicals
then combined with 2-acetylpyrrole (2) to form the radical
intermediate 3. Upon the voltage polarity reversal, 3 was
oxidized to the allylic cation 4 at the same electrode.
Subsequent deprotonation of 4 was rapid, generating the
final product 5. During ACE, the same chemical trans-
formations were also taking place on the other electrode.
During reaction optimization (Table 1), we varied the
frequency f and amplitude Vp of the square waveform. We
found that trifluoromethylation of 2 at 100 Hz and 4.4 V
(entry 5) afforded the desired product 5 in 84% yield. The
mono/bistrifluoromethylated product ratio (5/6) was found to
be 19:1. Both f and Vp strongly affected the yield and
selectivity. At 100 Hz, 5 was predominantly formed with a 5/6
ratio of >20:1 (entries 4 and 6). In contrast, 6 became more
favorable at 10 Hz (entry 7). At 1000 Hz, no reaction took
place (entry 8). If Vp was set greater or less than 4.4 V, the
yield of 5 decreased. At Vp < 4.4 V, there was a significant
amount of unreacted 2 recovered after 24 h (entries 2−4). At
Vp = 4.8 V, although 2 was completely consumed, a chlorinated
side product was isolated in 16% yield (entry 6). Importantly,
the use of 60 Hz sine waveform (entry 9), the same waveform
as the household power supply, afforded the desired product 5
in 40% NMR conversion. The application to a large scale was
explored utilizing 1 mmol of 2-acetylpyrrole (2), and the
B
Org. Lett. XXXX, XXX, XXX−XXX