Direct C(sp3)?H Trifluoromethylation of Unactivated Alkanes Enabled by Multifunctional Trifluoromethyl Copper Complexes

Article abstract of DOI:10.1002/anie.202012263

A mild and operationally simple C(sp³)?H trifluoromethylation method was developed for unactivated alkanes by utilizing a bench-stable Cu^III complex, bpyCu(CF₃)₃, as the initiator of the visible-light photoinduced reaction, the source of a trifluoromethyl radical as a hydrogen atom transfer reagent, and the source of a trifluoromethyl anion for functionalization. The reaction was initiated by the generation of reactive electrophilic carbon-centered CF₃ radical through photoinduced homolytic cleavage of bpyCu(CF₃)₃, followed by hydrogen abstraction from an unactivated C(sp³)?H bond. Comprehensive mechanistic investigations based on a combination of experimental and computational methods suggested that C?CF₃ bond formation was enabled by radical–polar crossover and ionic coupling between the resulting carbocation intermediate and the anionic CF₃ source. The methylene-selective reaction can be applied to the direct, late-stage trifluoromethylation of natural products and bioactive molecules.

Full text of DOI:10.1002/anie.202012263

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Direct C(sp3)–H Trifluoromethylation of Unactivated Alkanes
Enabled by Multifunctional Trifluoromethyl Copper Complexes
Authors: Geunho Choi, Geun Seok Lee, Beomsoon Park, Dongwook
Kim, and Soon Hyeok Hong
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202012263
Link to VoR: https://doi.org/10.1002/anie.202012263

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
Direct C(sp3)–H Trifluoromethylation of Unactivated Alkanes
Enabled by Multifunctional Trifluoromethyl Copper Complexes
Geunho Choi[a],[c], Geun Seok Lee[a],[c], Beomsoon Park[a], Dongwook Kim[a],[b], and Soon Hyeok Hong*[a]
[a]
G. Choi, G. S. Lee, B. Park, Dr. D. Kim, Prof. Dr. S. H. Hong
Department of Chemistry, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 34141 (Republic of Korea)
E-mail: soonhyeok.hong@kaist.ac.kr
[b]
[c]
Dr. D. Kim
Center for Catalytic Hydrocarbon Functionalizations, Institute for Basic Science (IBS), Daejeon 34141 (Republic of Korea)
G. Choi, G. S. Lee
Department of Chemistry, College of Natural Sciences, Seoul National University, Seoul 08826 (Republic of Korea)
Abstract: Despite the extensive development in trifluoromethylation,
introduction of a trifluoromethyl group into an unactivated C(sp3)–H
bond is significantly challenging. Herein, we report the development
of a mild and operationally simple C(sp3)–H trifluoromethylation
method for unactivated alkanes utilizing a bench-stable Cu(III)
complex, bpyCu(CF3)3, as a visible-light photo-induced reaction
initiator, a trifluoromethyl radical source as a hydrogen atom transfer
reagent, and a trifluoromethyl anion source for functionalization. The
reaction was initiated by reactive electrophilic carbon-centered CF3
radical generation from photo-induced homolytic cleavage of
bpyCu(CF3)3, which performs hydrogen abstraction from an
facilitate a streamlined synthesis of various trifluoromethylated
compounds, obviating the need for pre-functionalization. During
the preparation of this manuscript, MacMillan and co-workers
reported trifluoromethylation of aliphatic C(sp3)–H bonds
combining decatungstate photocatalysis and copper catalysis by
using the Togni reagent.[13] The elegant dual catalysis enabled the
direct C–CF3 bond formation at aliphatic and benzylic C(sp3)–H
bonds.
unactivated
investigations based on
C(sp3)–H
bond.
Comprehensive
mechanistic
a
combination of experimental and
computational methods suggested that C–CF3 bond formation was
enabled by radical-polar crossover and ionic coupling between the
resulting carbocation intermediate and the anionic CF3 source. The
newly developed reaction can be applied to various methylene
selective trifluoromethylation reactions including direct, late-stage
trifluoromethylation of natural products and bioactive molecules.
Introduction
The introduction of a trifluoromethyl group to organic
molecules is important for metabolic stability, lipophilicity, and
hydrophobicity. Therefore, trifluoromethylation has attracted
significant attention in the field of material, pharmaceutical, and
agricultural chemistry to conveniently modulate the polarity,
solubility, and chemical reactivity of desired compounds.[1]
Various synthetic methods to form C–CF3 bonds have been
developed using nucleophilic,[2] electrophilic,[3] or radical[4]
trifluoromethyl sources.
Direct formation of C–CF3 bonds via C–H activation is
ideal for trifluoromethylation in consideration of atom- and step-
efficiencies. C(sp2)–H trifluoromethylation reactions have been
extensively developed to introduce trifluoromethyl groups on
arenes,[5] heteroarenes,[6] olefins,[7] and aldehydes.[8] Although
C(sp2)–H trifluoromethylation has been significantly investigated,
very few examples of C(sp3)–H trifluoromethylation have been
reported, employing a directing group,[9] and being only applicable
to relatively more activated C(sp3)–H bonds such as α-
carbonyl,[10] α-nitrogen,[11] and benzylic C–H bonds (Scheme
1a).[12] Therefore, direct C(sp3)–H trifluoromethylation of
unactivated alkanes under mild conditions is highly desired to
Scheme 1. Undirected C(sp3)–CF3 trifluoromethylation via C(sp3)–H activation.
Considering
a
novel strategy for the direct
trifluoromethylation of normal C(sp3)–H bonds, we were inspired
by the recently highlighted bpyM(CF3)x (M = Cu, Zn; bpy = 2,2’-
bipyridine) complexes[8, 9b, 14] due to their exceptional reactivity
1
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
toward benzylic C(sp3)–CF3 trifluoromethylation (Scheme 1b). Li
and co-workers introduced copper-catalyzed benzylic C(sp3)–H
trifluoromethylation using bpyZn(CF3)2 as the anionic
trifluoromethyl source and N-fluorobenzenesulfonimide (NFSI) or
Selectfluor as the oxidant and hydrogen atom transfer (HAT)
reagent precursor.[12b] Liu and Cheng trifluoromethylated benzylic
C(sp3)–H bond by combining bpyCu(CF3)3 1a and Zn(Me)2 to
activate 1a to bpyCu(II)(CF3)2 with the generation of MeCF3.[12c]
Cook and co-workers reported the trifluoromethylation of benzylic
C(sp3)–H with 1a using persulfate, silane, and acid in
acetone/water.[12a] UV-irradiation induces homolysis of 1a to form
the active bpyCu(II)(CF3)2 and CF3 radical, while the sulfate
radical anion performs the HAT from the benzylic C–H bond,
generating the benzyl radical which is recombined with
bpyCu(II)(CF3)2, followed by reductive elimination, for the desired
trifluoromethylation. Especially, silane is required to capture the
CF3 radical to control the reactivity.
equiv of Oxone (39 or 70%, entries 7–8, respectively). Control
experiments in the dark demonstrated the critical role of irradiation
to produce 3a (entry 9). Altering the light source to the near
ultraviolet region (390 nm) was not effective (21%, entry 10). A
lower yield was obtained when the reaction was conducted under
air (63%, entry 11). Another oxidant including a persulfate salt,
(NH4)2S2O8, resulted in a lower yield (60%, entry 12). Other
solvents, except CH3CN, afforded lower product yield (Table S1).
The conditions reported by Cook were not effective to afford 3a in
good yield (14%, entry 13) likely because the HAT from silane to
the generated sulfate radical anion is faster than that from a
C(sp3)–H bond of 1a, which has a higher BDE than a benzylic
C(sp3)–H bond.[12a]
Table 1. Variation of the reaction conditions.[a]
The ingenious strategy developed by Cook led us to
consider a different strategy to functionalize unactivated C(sp3)–
H bonds in order to expand the scope from benzylic C(sp3)–H
bonds. Instead of using an externally added persulfate as the HAT
reagent and trapping the reactive CF3 radical with a silane to
control the reactivity, we envisioned direct utilization of the highly
reactive CF3 radical without using an external HAT reagent.
Based on the bond-dissociation energy (BDE) of fluoroform (106
kcal/mol),[15] the CF3 radical is hypothesized to perform the HAT
from unactivated C(sp3)–H bonds of alkanes to generate alkyl
radicals, which can further undergo Cu-complex-mediated
trifluoromethylation. Although the CF3 radical has been reported
to perform the HAT from C–H bonds of various hydrocarbons to
generate fluoroform,[16] to the best of our knowledge, no C–H
functionalization reaction utilizing the CF3 radical as the HAT
reagent has been reported. Herein, a novel trifluoromethylation of
unactivated C(sp3)–H bonds was achieved by photo-induced
bpyCu(CF3)3 1a based on the newly designed strategy (Scheme
1c). A series of control experiments and computational studies
suggested that the reaction occurs via an ionic pathway enabled
by an oxidative radical-polar crossover with the aid of an Oxone
additive. Notably, the bench-stable Cu complex 1a plays multiple
roles of the photo-induced reaction initiator, trifluoromethyl radical
source for HAT, and trifluoromethyl anion source, thus enabling
direct trifluoromethylation of strong C(sp3)–H bonds under mild
conditions.
Entry
1[b]
2
Variation from the “standard conditions”
None
Yield 3a [%]
93
1.0 equiv 2a
50
3
3.0 equiv 2a
73
4
16.0 equiv 2a
>96
>96
16
5
32.0 equiv 2a
6[c]
7
No Oxone
1.0 equiv Oxone
2.0 equiv Oxone
No light
39
8
70
9
N. D.
21
10
11
12
13[d]
390 nm, instead of blue light
Under air in a closed vial
(NH4)2S2O8, instead of Oxone
Cook conditions[12a]
63
60
14
[a] Yields were determined by GC analysis using hexafluorobenzene as an
internal standard. [b] Reaction conditions: 2a (0.25 mmol, 5.0 equiv), 1a (0.050
mmol, 1.0 equiv), Oxone (0.15 mmol, 3.0 equiv), and CH3CN (0.5 mL) under 40
W blue LED irradiation with fan cooling (30±5 °C) for 3 h. [c] 12 h. [d] Reaction
conditions: 2a (0.10 mmol, 2.0 equiv), 1a (0.05 mmol, 1.0 equiv), (NH4)2S2O8
(0.15 mmol, 3.0 equiv), iPr3SiH (0.15 mmol, 3.0 equiv), trifluoroacetic acid (0.40
mmol, 8.0 equiv), acetone (0.3 mL), and water (0.3 mL) under 43 W 370 nm
LED irradiation with fan cooling (30±5 °C) for 18 h. N. D. = not detected.
Results and Discussion
Cyclohexane 2a was chosen as a model hydrocarbon
substrate to evaluate the feasibility of the designed strategy with
1a (Table 1). When 2a (5.0 equiv) was reacted with 1a (1.0 equiv)
in the presence of Oxone (3.0 equiv) in CH3CN (0.5 mL),
(trifluoromethyl)cyclohexane 3a was produced in 93% yield under
blue light irradiation for 3 h (entry 1). Reducing the amount of 2a
(1 or 3 equiv) resulted in diminished yield (50% and 73%, entries
2–3, respectively), and further increasing 2a content showed
quantitative yield (>96%, entries 4–5). Decreased yield was
afforded without Oxone (16%, entry 6) and with less than 3.0
Under optimized conditions, the reactivity of various
Cu(III)–CF3 complexes bearing other N,N-bidentate ligands,
bipyridines, and phenanthrolines, were investigated (Table 2).
Among the bipyridine ligands tested, the simplest bipyridine
performed best. Introduction of any substituent on the ortho, meta,
or para positions to modify the steric and electronic characters of
the ligand resulted in lower product yield (1b–1h). No clear
correlation between electronic and steric characters of the ligand
2
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
and reaction outcome were observed. The absorption of the
Cu(III) complexes was also examined, but no specific correlations
between the absorptions and λmax of the screened Cu complexes
and the reaction yields were observed (Figure S1–5).
conditions with water, the yield was dramatically increased to 84%,
although 10% of cycloheptanone 4c was observed as the by-
product. Cyclic alkanes with varied ring sizes were efficiently
converted to the corresponding trifluoromethylated products (3a–
3e). The reactions with some substrates (3e, 3j, 3l, 3m, 3q, 3u)
proceeded smoothly in hexafluoroisopropanol (HFIP) instead of
CH3CN. For example, cyclododecane 3e furnished the desired
product in 30% yield when HFIP was used as the solvent,
whereas unidentified complex mixtures were obtained in CH3CN
solvent. Linear alkanes, including hexane (2f) and pentane (2g),
readily underwent trifluoromethylation in very good yields with a
preponderance on methylene positions (76 and 87% yields; 88
and 85% methylene selectivity, respectively), likely due to the
enhanced stability of methylene radicals compared to primary
methyl radicals. Benzylic C–H bonds in toluene could also be
functionalized, but in a low yield (39%, 3h) due to competing CF3
radical addition to the aryl ring, which was suppressed when
Table 2. Reactivity of various copper complexes.[a]
silane was used as
a trapping reagent under Cook’s
conditions.[12a] Increased steric hindrance around the arene ring
inhibited the undesired addition, affording the desired product in
excellent yield (91%, 3i). Carboxylic acids (2j and 2k), an ester
(2l), and an amide (2m) were also tolerated in this transformation.
1-Propanoic acid (2j) and methyl propanoate (2l) mostly
underwent α-methylene C–H trifluoromethylation (47 and 71%
yields; 85 and 82% α-selectivity, respectively), consistently
demonstrating preference for the methylene C–H bond over
terminal methyl C–H bond. Propanamide (2m) afforded the α-
methylene C–H trifluoromethylated product (37%, 3m) without
significant byproduct formation. With 1-butanoic acid (2k), the
more hydric β-methylene position was preferred over the α-
methylene C–H bond (62% yield; 62% β-selectivity, 3k). Other
hydrocarbons
containing
with
a
carbonyl
group
yields,
were
trifluoromethylated
satisfactory
favoring
[a] Reaction conditions: 2a (0.25 mmol, 5.0 equiv), 1 (0.050 mmol, 1.0 equiv),
Oxone (0.15 mmol, 3.0 equiv), and CH3CN (0.5 mL) under 40 W blue LED
irradiation with fan cooling (30±5 °C) for 12 h; Yields were determined by GC
analysis using hexafluorobenzene as an internal standard.
functionalization on the more electron-rich C–H bonds due to the
preference of the electrophilic CF3 radical for HAT (3n–3q). Cyclic
ketones including cyclopentanone (2n) and cyclohexanone (2o)
afforded trifluoromethylated products favoring the β-methylene
C–H position (58 and 80% yields; 79 and 73% β-selectivity,
respectively). Especially, the linear carbonyl 4-heptanone (2p)
was exclusively trifluoromethylated at the β-position of the
carbonyl group rather than the α-methylene or ⋎-methyl position
(88%, 3p). The β-ketoester was also compatible with a major β-
trifluoromethylation on the carbonyl group (48%, 3q).
Unfortunately, reactions using aliphatic amines such as N-Boc-
pyrrolidine and N-acetylpyrrolidine were not compatible,
producing messy unidentifiable compounds.
After investigating the reaction conditions, various
hydrocarbon substrates were investigated to evaluate the
generality of the newly developed photo-induced C–H
trifluoromethylation protocol (Table 3). Although the conditions
using anhydrous acetonitrile solvent performed best for 2a, co-
solvent conditions with water (CH3CN/H2O, 5:1 (v/v)) using 6.0
equiv hydrocarbon substrates were more effective for other C–H
substrates, presumably due to increased solubility of Oxone. For
instance, 2c produced (trifluoromethyl)cycloheptane 3c in 10%
yield under the conditions without water. With the co-solvent
3
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
Table 3. Substrate scope of unactivated C(sp3)–H trifluoromethylation.[a]
[a] Reaction conditions: 2 (0.60 mmol, 6.0 equiv), 1a (0.10 mmol, 1.0 equiv), Oxone (0.30 mmol, 3.0 equiv), CH3CN (1.0 mL), and H2O (0.2 mL) under 40 W blue
LED irradiation with fan cooling (30±5 °C) for 3 h; Yields and selectivities (%) were determined by 19F NMR using fluorobenzene as an internal standard unless
otherwise noted. [b] Reaction conditions: 2a (0.25 mmol, 5.0 equiv), 1a (0.050 mmol, 1.0 equiv), Oxone (0.15 mmol, 3.0 equiv), and CH3CN (0.5 mL) under 40 W
blue LED irradiation with fan cooling (30±5 °C) for 3 h; Yields were determined by GC analysis using hexafluorobenzene as internal standard. [c] HFIP (2.0 mL) as
a solvent. [d] Isolated yield. [e] 2w (0.30 mmol, 3.0 equiv).
4
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
Trifluoromethyl group introduction can reportedly
regulate bioactivity.[1b] Previously studied late-stage C–H
trifluoromethylation mainly focused on C(sp2)–H[6a, 17] and weak
noted that no DCF3 was detected in the reaction with 2a in CD3CN
indicating the reaction between a trifluoromethyl source and
CD3CN is not facile. Without cyclohexane, DCF3 was not
observed from the reaction between 1a and CD3CN, confirming
that hydrogen or deuterium in the generated fluoroforms mainly
originated from cyclohexane (entry 3). When the reaction was
performed with 2a-d12 in CH3CN, DCF3 (>8%) was observed
along with a non-negligible amount of HCF3 (>12%, entry 4). This
may originate from HAT from CH3CN because the reaction
between 1a and CH3CN without cyclohexane generated HCF3
(>40%, entry 5). These results clearly indicate that the
trifluoromethyl radical serves as the HAT reagent in the reaction.
12b, 13]
C(sp3)–H bonds.[12a,
The applicability of the newly
developed reaction was further examined with direct C(sp3)–H
trifluoromethylations of well-known natural products and bioactive
molecule derivatives (Table 3). Eucalyptol, which has anti-
inflammatory and antioxidant effects,[18] was smoothly
trifluoromethylated with high regio- and diastereo-selectivities at
the less sterically hindered position (41%, 3r). Other bicyclic
terpenoids such as norcamphor, the analog of the naturally
occurring terpenoid camphor, and fenchone, which exhibits
antifungal activity,[19] also delivered the desired C–H
trifluoromethylated products (47%, 3s and 31%, 3t). Additionally,
Table 4. Control experiments regarding the CF3 radical as a HAT reagent.
a
cubane core, whose derivatives have not been
trifluoromethylated, afforded 3u in moderate yield (14%).
Ambroxide, a naturally occurring terpenoid that is commonly used
in the fragrance industry,[20] was trifluoromethylated to furnish a
major 2-functionalized regioisomer albeit with low efficiency (20%,
3v). It should be noted that trifluoromethylation was conducted in
the sterically less-hindered cyclohexyl ring, but not at the α-oxy
position, exhibiting an unusual trend compared to other radical-
based C–H functionalization reactions.[21] Sclareolide, an
antifungal plant compound,[22] showed a similar selectivity to
ambroxide with higher yield (38%, 3w). Finally, camphoric acid,
which is often used as a precursor of bioactive molecules,[23] was
used as a dimethyl ester derivative (2x) and was selectively
trifluoromethylated at the least sterically hindered methylene
position in
a moderate yield (32%, 3x). These results
Entry
2a or 2a-d12
Solvent
3a or 3a-d11
HCF3
[%]
DCF3
[%]
demonstrated the potential applicability of the developed reaction
for late-stage, direct C–H trifluoromethylation of functional
molecules with reasonable regio- and diastereo-selectivities
favoring less-sterically hindered positions.
[%]
1
2
3
4
5
2a
CD3CN
CD3CN
CD3CN
CH3CN
CH3CN
35
17
-
>34
>9
N. D.
>12
N. D.
>8
2a-d12
To determine the underlying reaction mechanism,
several control experiments and computational studies were
conducted. The intermediary of the CF3 radical in the reaction was
first examined (Table 4a). The liberation of the CF3 radical from
1a is experimentally and computationally well-documented.[12a, 14a,
-
2a-d12
-
>5
12
-
>12
>40
-
14b]
[a] Yields determined by GC analysis using hexafluorobenzene as an internal
standard. [b] Yields determined by 19F NMR analysis using fluorobenzene as an
internal standard. [c] Product was confirmed by HRMS. [d] NMR tube reactions
were performed. The amount of dissolved gaseous fluoroform was only
determined by 19F NMR analysis using fluorobenzene as an internal standard.
[e] 2a-d12 (99.5 D atom%); CD3CN (99.8 D atom%). N. D. = not detected.
In agreement with the previous reports, introduction of a
radical scavenger, TEMPO (2,2,6,6-tetramethylpiperidine-N-oxyl
radical) or BHT (2,6-di-tert-butyl-4-methylphenol), suppressed the
trifluoromethylation of 2a, with simultaneous formation of the
TEMPO–CF3 adduct in 94% yield, clearly indicating that CF3
radical formation is involved in the product-forming pathway.
After identification of the CF3 radical as the HAT reagent,
three product forming pathways were proposed; (1) direct
coupling between the alkyl and CF3 radicals, (2) C–CF3 reductive
elimination via a Cu(III)–alkyl complex similar to recent reports on
1a-mediated trifluoromethylation reactions,[12a, 12c, 14a, 14c] and (3)
radical-polar crossover involving ionic coupling between an alkyl
cation and trifluoromethyl anion. The first possibility, involving the
selective coupling of two transient radical species, is unlikely
because no homodimerization product was observed with or
without Oxone. This indicates that the free alkyl radical species
may not be sufficiently generated for the radical-radical coupling
reaction. Secondly, reductive elimination is a viable pathway for
C–CF3 bond formation with 1a as the trifluoromethylating reagent.
Next, deuterium labelling experiments using deuterated
cyclohexane (2a-d12) and acetonitrile were conducted to
determine the applicability of the generated CF3 radical as the
HAT reagent against the aliphatic C–H bonds (Table 4b). A sulfate
radical, which can be generated from Oxone additive, was also
reported as a HAT reagent that can activate C–H bonds.[8, 12a, 12c]
However, as the reaction proceeded without Oxone, albeit in a
lower yield (16%, Table 1, entry 6), it is unlikely that the HAT
mediated by the sulfate radical is the major C–H activating
process. With 2a and 2a-d12 as C–H substrates in a CD3CN
solvent, fluoroforms (HCF3, >34% and DCF3, >12%) were
observed by 19F NMR spectroscopy (entries 1–2). It should be
5
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
A Cu(II) complex has been proposed to trap a carbon-centered
radical, subsequently forming the desired C–CF3 bond via
concerted reductive elimination.[12a, 12c, 14a, 14c] Recently, anionic or
neutral alkyl-CuIII(CF3)x complexes were investigated to disclose
that C(sp3)–CF3 bonds can be formed via concerted reductive
elimination from a high-valent copper species.[24] Lastly, a radical-
polar crossover pathway for the ionic coupling was proposed. In
this case, alkyl radical, which has low oxidation potential
(Eocalc(Cy-rad/Cy-cat) = +0.46 V), is oxidized to generate an alkyl
cation, followed by ionic coupling with a CF3 anion for bond
formation. C–X (X = F, CF3) bonds can be formed via electron
transfer from the alkyl radical to CuII–X.[14d, 25] We hypothesized
that the oxidant can accelerate electron transfer from the alkyl
radical, facilitating the radical-polar crossover pathway.
density functional theory (TD-DFT) computations on complex 1a
revealed relevant singlet transitions in the blue light region, giving
rise to 1a* (Figure S15). This complex would then undergo
intersystem crossing (ISC), traversing the triplet excited 31a to
release CF3 radical CF3-rad and 2I, as previously reported by the
Cook group.[14a] Beginning with this species, HAT of CF3-rad
against cyclohexane 2a is facile with a barrier of only 10.9
kcal/mol. This is reasonable considering the relative instability of
the CF3 radical and its electrophilic nature. After the HAT event,
the reaction pathway could follow one of the two aforementioned
mechanisms. The reductive elimination pathway (orange trace)
involving the Cy–Cu(III) complex 1II exhibited an overall barrier of
20.8 kcal/mol. All attempts to find alternative pathways of
1
reductive elimination from II with a lower energy barrier which
can compete with the oxidative pathway (vide infra) were
unsuccessful. The bisulfate ligand-assisted reductive elimination
process, as proposed by the Cook group, has a computed
activation barrier of 14.8 kcal/mol (Figure S14).[14a]
To determine the most viable pathway, density-
functional theory (DFT) studies of the two proposed pathways
were conducted with 2a as model substrate at the B3LYP-D3/6-
311++G**/SDD level of theory (Figure 1). Time-dependent
Figure 1. Computational results following the proposed reaction pathway.
6
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
Scheme 2. H/D kinetic isotope effect.
In contrast, the DFT computation suggested that
chemical oxidation of 2I with Oxone affords a Cu(III)–bisulfate
species 1IV (blue trace), located at –35.1 kcal/mol. This is due to
the strong oxidation power of Oxone, providing results analogous
to previous work by the Cook group where persulfate served as
the oxidant to produce the identical Cu species (E°calc[2I/1IV] =
+0.98 V, E°(Oxone) = +1.81 V).[14a] With this Cu species, single-
electron oxidation of cyclohexyl radical was modeled using the
Marcus theory, and the energy barrier of the electron transfer was
5.3 kcal/mol. The exotherm of the oxidation was –5.0 kcal/mol,
indicating a good match in terms of redox potentials (E°calc[1IV/2IV-
anion] = +0.67 V, E°calc(Cy-rad/Cy-cat) = +0.46 V). The
generated transient Cu(II) complex 2IV-anion will undergo
structural rearrangement to liberate a trifluoromethyl anion, which
can spontaneously recombine with the cyclohexyl cation Cy-cat
to furnish the desired product with a strong driving force of 67.8
kcal/mol. It should be noted that the oxidation of other Cu species,
including 1II, could eventually generate the 2IV-anion and Cy-cat.
Overall, for unactivated alkanes, the activation barrier was much
lower for the radical-polar crossover pathway (5.3 kcal/mol)
compared to the reductive elimination pathway (20.8 kcal/mol),
accounting for the improved reaction efficiency with an oxidant. At
the current stage, we cannot rule out another pathway leading to
Secondly, we attempted to trap an alkyl cation
intermediate that could arise from the radical-polar crossover
pathway (Table 5). From Table 1, the standard reaction with 2a
under anhydrous conditions produced 3a without significant by-
product production (entry 1). Upon water addition, a significant
amount of cyclohexanone 4a (14%) was observed, which can be
generated from the reaction between the cyclohexyl cation and
water to furnish cyclohexanol, followed by subsequent oxidation
to form the ketone (entry 2). Increased water contents facilitated
the formation of 4a (20%, entry 3). Without Oxone, no 4a was
generated despite the existence of water (entry 4). Cycloheptane
2c also produced 3c in a decreased yield (25%) with increased
cycloheptanone by-product 4c (16%) under the co-solvent
conditions of increased amount of water (CH3CN/H2O, 1:1 (v/v)).
This clearly suggested that a carbocation is generated under the
reaction conditions. It has been reported that a persulfate can be
homolytically cleaved via UV irradiation, acting as a HAT
reagent.[12a. 26] However, under the reaction conditions without 1a,
neither oxidation nor dimerized products were observed,
indicating that Oxone itself cannot perform the HAT from 2a under
the blue light irradiation conditions used herein (entry 5).
2
the key intermediates, IV-anion and Cy-cat, which is the direct
In summary, driven by the photo irradiation of 1a, the
homolytically cleaved CF3 radical conducts the HAT from a
C(sp3)–H bond of alkane substrate, as experimentally supported
by trapping CF3 radical and detection of fluoroform HCF3. Oxone
dramatically increased the reaction efficiency by oxidizing the
resulting Cu(II) complex 2I to Cu(III) 1IV which can oxidize the alkyl
radical to form an alkyl cation. Formation of 4a in presence of
water supports the existence of an alkyl cation. The produced
anionic Cu(II) complex 2IV-anion acts as a CF3 anion source that
undergoes ionic coupling with the alkyl cation to form the C(sp3)–
CF3 bond.
single electron transfer between Cy-rad and Oxone to produce
Cy-cat. Subsequent association of the bisulfate anion may lead
to the formation of the identical key intermediates (blue trace via
2I + Cy-cat).
Based on the computational study, further experiments
were performed to verify the proposed mechanism. First, the
kinetic isotope effect (KIE) was measured using 2a and 2a-d12 to
elucidate the rate-determining step (Scheme 2). Primary KIE
values of 4.2 (parallel reaction) and 5.2 (intermolecular
competition) were obtained, clearly indicating that the HAT step is
rate-determining, in good accordance with the DFT studies. If the
reductive elimination were the major product forming pathway, a
primary KIE would not be observed as the reductive elimination
should be the rate-determining step (Figure 1). Besides, as the
only masses of the products obtained from the intermolecular
competition experiment were m/z 152.1 (3a) and 163.2 (3a-d11),
we could conclude that the C–H cleavage is irreversible. This
phenomenon is in good agreement with our DFT-computed
energy profile where reverse hydrogen atom transfer possesses
a higher barrier compared to the forward reactions.
Table 5. Oxidation product detection.[a]
Entry
Variation from the “standard conditions”
Yield 3a
Yield 4a
[%]
[%]
1
2
3
4
5
No water
93
63
19
31
-
N. D.
14
None
CH3CN/H2O (1:1 (v/v))
No Oxone
No 1a
20
N. D.
N. D.
[a] Yields determined by GC analysis using hexafluorobenzene as an internal
standard.
7
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
Kuninobu, M. Kanai, Nat. Commun. 2014, 5, 3387; d) Y. Fujiwara,
J. A. Dixon, F. O'Hara, E. D. Funder, D. D. Dixon, R. A. Rodriguez,
R. D. Baxter, B. Herlé, N. Sach, M. R. Collins, Y. Ishihara, P. S.
Baran, Nature 2012, 492, 95-99; e) L. Chu, F.-L. Qing, J. Am. Chem.
Soc. 2012, 134, 1298-1304; f) M. Nagase, Y. Kuninobu, M. Kanai,
J. Am. Chem. Soc. 2016, 138, 6103-6106.
a) C. Feng, T.-P. Loh, Angew. Chem. Int. Ed. 2013, 52, 12414-
12417; Angew. Chem. 2013, 125, 12640-12643; b) N. J. W.
Straathof, S. E. Cramer, V. Hessel, T. Noël, Angew. Chem. Int. Ed.
2016, 55, 15549-15553; Angew. Chem. 2016, 128, 15778-15782; c)
C. Yang, A. Hassanpour, K. Ghorbanpour, S. Abdolmohammadi, E.
Vessally, RSC Adv. 2019, 9, 27625-27639.
P. Zhang, H. Shen, L. Zhu, W. Cao, C. Li, Org. Lett. 2018, 20, 7062-
7065.
a) T. Ide, S. Masuda, Y. Kawato, H. Egami, Y. Hamashima, Org.
Lett. 2017, 19, 4452-4455; b) Z. Liu, H. Xiao, B. Zhang, H. Shen, L.
Zhu, C. Li, Angew. Chem. Int. Ed. 2019, 58, 2510-2513; Angew.
Chem. 2019, 131, 2532-2535.
a) A. E. Allen, D. W. C. MacMillan, J. Am. Chem. Soc. 2010, 132,
4986-4987; b) A. T. Herrmann, L. L. Smith, A. Zakarian, J. Am.
Chem. Soc. 2012, 134, 6976-6979; c) D. A. Nagib, M. E. Scott, D.
W. C. MacMillan, J. Am. Chem. Soc. 2009, 131, 10875-10877; d)
Q.-H. Deng, H. Wadepohl, L. H. Gade, J. Am. Chem. Soc. 2012,
134, 10769-10772; e) V. Matoušek, A. Togni, V. Bizet, D. Cahard,
Org. Lett. 2011, 13, 5762-5765.
a) A. A. S. Gietter-Burch, V. Devannah, D. A. Watson, Org. Lett.
2017, 19, 2957-2960; b) L. Chu, F.-L. Qing, Chem. Commun. 2010,
46, 6285-6287; c) H. Mitsudera, C.-J. Li, Tetrahedron Lett. 2011, 52,
1898-1900.
a) S. Guo, D. I. AbuSalim, S. P. Cook, J. Am. Chem. Soc. 2018, 140,
12378-12382; b) H. Xiao, Z. Liu, H. Shen, B. Zhang, L. Zhu, C. Li,
Chem 2019, 5, 940-949; c) M. Paeth, W. Carson, J.-H. Luo, D.
Tierney, Z. Cao, M.-J. Cheng, W. Liu, Chem. Eur. J. 2018, 24,
11559-11563; d) H. Egami, T. Ide, Y. Kawato, Y. Hamashima, Chem.
Commun. 2015, 51, 16675-16678.
P. J. Sarver, V. Bacauanu, D. M. Schultz, D. A. DiRocco, Y.-H. Lam,
E. C. Sherer, D. W. C. MacMillan, Nat. Chem. 2020, 12, 459-467.
a) S. Guo, D. I. AbuSalim, S. P. Cook, Angew. Chem. Int. Ed. 2019,
58, 11704-11708; Angew. Chem. 2019, 131, 11830-11834; b) H.
Shen, Z. Liu, P. Zhang, X. Tan, Z. Zhang, C. Li, J. Am. Chem. Soc.
2017, 139, 9843-9846; c) X. Tan, Z. Liu, H. Shen, P. Zhang, Z.
Zhang, C. Li, J. Am. Chem. Soc. 2017, 139, 12430-12433; d) H.
Xiao, H. Shen, L. Zhu, C. Li, J. Am. Chem. Soc. 2019, 141, 11440-
11445; e) A. M. Romine, N. Nebra, A. I. Konovalov, E. Martin, J.
Benet-Buchholz, V. V. Grushin, Angew. Chem. Int. Ed. 2015, 54,
2745-2749; Angew. Chem. 2015, 127, 2783-2787.
Conclusion
In conclusion, a novel C(sp3)–H trifluoromethylation was
developed using a photo-induced high-valent Cu–CF3 complex.
Diverse unactivated alkanes including bioactive molecules were
trifluoromethylated using bench-stable bpyCu(CF3)3 1a under
mild reaction conditions, favoring the methylene and less-
sterically hindered C(sp3)–H bonds. The experimental and
computational mechanistic studies suggested that the reaction
proceeds via the CF3 radical-mediated HAT reaction to activate
C(sp3)–H bonds, followed by radical-polar crossover and ionic
coupling. Notably, 1a performs multiple roles as the photo-
induced reaction initiator, precursor of the CF3 radical as a unique
HAT reagent, and trifluoromethylating source. It is anticipated that
the developed operatively simple reaction will have wide-scale
[7]
[8]
[9]
[10]
applications
especially
for
late-stage,
single-step
trifluoromethylation of functional molecules.
[11]
[12]
Acknowledgements
Dr. Kicheol Kim is gratefully acknowledged for preliminary studies.
We thank Sehye Min, Jinwoo Kim, and Jeonguk Kweon for the
help with the spectroscopic measurements. This work was
supported by the National Research Foundation of Korea (NRF)
[13]
[14]
grant
funded
by
the
Korea
government
(NRF-
2019R1A2C2086875; NRF-2014R1A5A1011165, Center for New
Directions in Organic Synthesis).
Conflict of interest
[15]
[16]
J. C. Amphlett, J. W. Coomber, E. Whittle, J. Phys. Chem. 1966, 70,
593-594.
The authors declare no conflict of interest.
a) P. Gray, A. A. Herod, A. Jones, Chem. Rev. 1971, 71, 247-294;
b) G. Bullock, R. Cooper, Trans. Faraday Soc. 1970, 66, 2055-2064;
c) T. N. Bell, U. F. Zucker, Can. J. Chem. 1970, 48, 1209-1213; d)
R. A. Fass, J. E. Willard, J. Chem. Phys. 1970, 52, 1874-1882.
F. Ye, F. Berger, H. Jia, J. Ford, A. Wortman, J. Börgel, C. Genicot,
T. Ritter, Angew. Chem. Int. Ed. 2019, 58, 14615-14619; Angew.
Chem. 2019, 131, 14757-14761.
G. H. Seol, K. Y. Kim, in Drug Discovery from Mother Nature (Eds.:
S. C. Gupta, S. Prasad, B. B. Aggarwal), Springer International
Publishing, Cham, 2016, pp. 389-398.
S. Sukhchain, G. Pawan, G. Jeena, Curr. Drug Discov. Technol.
2019, 16, 1-12.
P. Zerbe, J. Bohlmann, in Biotechnology of Isoprenoids (Eds.: J.
Schrader, J. Bohlmann), Springer International Publishing, Cham,
2015, pp. 427-447.
a) C. M. Morton, Q. Zhu, H. Ripberger, L. Troian-Gautier, Z. S. D.
Toa, R. R. Knowles, E. J. Alexanian, J. Am. Chem. Soc. 2019, 141,
13253-13260; b) S. Mukherjee, B. Maji, A. Tlahuext-Aca, F. Glorius,
J. Am. Chem. Soc. 2016, 138, 16200-16203.
M. Jasiński, Y. Stukkens, H. Degand, B. Purnelle, J. Marchand-
Brynaert, M. Boutry, Plant Cell 2001, 13, 1095-1107.
V. V. Chernyshov, O. I. Yarovaya, D. S. Fadeev, Y. V. Gatilov, Y. L.
Esaulkova, A. S. Muryleva, K. O. Sinegubova, V. V. Zarubaev, N. F.
Salakhutdinov, Mol. Divers. 2020, 24, 61-67.
a) M. Paeth, S. B. Tyndall, L.-Y. Chen, J.-C. Hong, W. P. Carson, X.
Liu, X. Sun, J. Liu, K. Yang, E. M. Hale, D. L. Tierney, B. Liu, Z. Cao,
M.-J. Cheng, W. A. Goddard, 3rd, W. Liu, J. Am. Chem. Soc. 2019,
141, 3153-3159; b) S. Liu, H. Liu, S. Liu, Z. Lu, C. Lu, X. Leng, Y.
Lan, Q. Shen, J. Am. Chem. Soc. 2020, 142, 9785-9791.
Z. Liu, H. Chen, Y. Lv, X. Tan, H. Shen, H.-Z. Yu, C. Li, J. Am. Chem.
Soc. 2018, 140, 6169-6175.
Keywords: C–H activation • copper • late-stage functionalization
• radical-polar crossover • trifluoromethylation
[17]
[18]
[1]
a) T. Fujiwara, D. O’Hagan, J. Fluor. Chem. 2014, 167, 16-29; b) N.
A. Meanwell, J. Med. Chem. 2018, 61, 5822-5880; c) Y. Zhao, M. G.
Schwab, A. Kiersnowski, W. Pisula, M. Baumgarten, L. Chen, K.
Müllen, C. Li, J. Mater. Chem. C 2016, 4, 4640-4646.
[19]
[20]
[2]
[3]
a) X. Liu, C. Xu, M. Wang, Q. Liu, Chem. Rev. 2015, 115, 683-730;
b) L. Chu, F.-L. Qing, Acc. Chem. Res. 2014, 47, 1513-1522.
a) A. Prieto, O. Baudoin, D. Bouyssi, N. Monteiro, Chem. Commun.
2016, 52, 869-881; b) S. Barata-Vallejo, B. Lantaño, A. Postigo,
Chem. Eur. J. 2014, 20, 16806-16829; c) J. Charpentier, N. Früh, A.
Togni, Chem. Rev. 2015, 115, 650-682; d) R. Tomita, T. Koike, M.
Akita, Angew. Chem. Int. Ed. 2015, 54, 12923-12927; Angew. Chem.
2015, 127, 13115-13119.
[21]
[4]
[5]
a) T. Koike, M. Akita, Acc. Chem. Res. 2016, 49, 1937-1945; b) A.
Studer, Angew. Chem. Int. Ed. 2012, 51, 8950-8958; Angew. Chem.
2012, 124, 9082-9090; c) A. Sato, J. Han, T. Ono, A. Wzorek, J. L.
Aceña, V. A. Soloshonok, Chem. Commun. 2015, 51, 5967-5970.
a) F. Sladojevich, E. McNeill, J. Börgel, S.-L. Zheng, T. Ritter,
Angew. Chem. Int. Ed. 2015, 54, 3712-3716; Angew. Chem. 2015,
127, 3783-3787; b) X.-G. Zhang, H.-X. Dai, M. Wasa, J.-Q. Yu, J.
Am. Chem. Soc. 2012, 134, 11948-11951; c) X. Wang, L. Truesdale,
J.-Q. Yu, J. Am. Chem. Soc. 2010, 132, 3648-3649; d) N. D. Litvinas,
P. S. Fier, J. F. Hartwig, Angew. Chem. Int. Ed. 2012, 51, 536-539;
Angew. Chem. 2012, 124, 551-554; e) M. Shang, S.-Z. Sun, H.-L.
Wang, B. N. Laforteza, H.-X. Dai, J.-Q. Yu, Angew. Chem. Int. Ed.
2014, 53, 10439-10442; Angew. Chem. 2014, 126, 10607-10610; f)
L. Li, X. Mu, W. Liu, Y. Wang, Z. Mi, C.-J. Li, J. Am. Chem. Soc.
2016, 138, 5809-5812.
[22]
[23]
[24]
[25]
[26]
a) Y. Liu, Y. Wang, Y. Yin, J. Pan, J. Zhang, Energy Fuels 2018, 32,
1999-2008; b) Y. Wang, Y. Wang, Y. Liu, Chem. Eng. J. 2020, 393,
124740.
[6]
a) D. A. Nagib, D. W. C. MacMillan, Nature 2011, 480, 224-228; b)
A. G. O'Brien, A. Maruyama, Y. Inokuma, M. Fujita, P. S. Baran, D.
G. Blackmond, Angew. Chem. Int. Ed. 2014, 53, 11868-11871;
Angew. Chem. 2014, 126, 12062-12065; c) T. Nishida, H. Ida, Y.
8
This article is protected by copyright. All rights reserved.

10.1002/anie.202012263
Angewandte Chemie International Edition
RESEARCH ARTICLE
Table of Contents
A photo-induced C(sp3)–H trifluoromethylation of alkanes is developed by employing bpyCu(CF3)3 as a multifunctional reagent; photo-
induced reaction initiator, the precursor of hydrogen atom transfer reagent, and trifluoromethylating source. The operatively simple
reaction enables the direct, late-stage trifluoromethylation of complex molecules under mild reaction conditions.
9
This article is protected by copyright. All rights reserved.