Mechanistic study of visible light-driven CdS or g-C3N4-catalyzed C–H direct trifluoromethylation of (hetero)arenes using CF3SO2Na as the trifluoromethyl source

Article abstract of DOI:10.1016/j.jcat.2020.06.032

The mild and sustainable methods for C–H direct trifluoromethylation of (hetero)arenes without any base or strong oxidants are in extremely high demand. Here, we report that the photo-generated electron-hole pairs of classical semiconductors (CdS or g-C₃N₄) under visible light excitation are effective to drive C–H trifluoromethylation of (hetero)arenes with stable and inexpensive CF₃SO₂Na as the trifluoromethyl (TFM) source via radical pathway. Either CdS or g-C₃N₄ propagated reaction can efficiently transform CF₃SO₂Na to ^[rad]CF₃ radical and further afford the desired benzotrifluoride derivatives in moderate to good yields. After visible light initiated photocatalytic process, the key elements (such as F, S and C) derived from the starting TFM source of CF₃SO₂Na exhibited differential chemical forms as compared to those in other oxidative reactions. The photogenerated electron was trapped by chemisorbed O₂ on photocatalysts to form superoxide radical anion (O₂^[rad]?) which will further attack ^[rad]CF₃ radical with the generation of inorganic product F^? and CO₂. This resulted in a low utilization efficiency of ^[rad]CF₃ (<50percent). When nitro aromatic compounds and CF₃SO₂Na served as the starting materials in inert atmosphere, the photoexcited electrons can be directed to reduce the nitro group to amino group rather than being trapped by O₂. Meanwhile, the photogenerated holes oxidize SO₂CF₃^? into ^[rad]CF₃. Both the photogenerated electrons and holes were engaged in reductive and oxidative paths, respectively. The desired product, trifluoromethylated aniline, was obtained successfully via one-pot free-radical synthesis.

Full text of DOI:10.1016/j.jcat.2020.06.032

Journal Pre-proofs
Mechanistic study of visible light-driven CdS or g-C3N4-catalyzed C-H direct
trifluoromethylation of (hetero)arenes using CF3SO2Na as the trifluoromethyl
source
Lele Wang, Ming Liu, Wenying Zha, Yingcong Wei, Xiongfeng Ma,
Chunwang Xu, Chenggang Lu, Nanfang Qin, Li Gao, Wenzhao Qiu, Rongjian
Sa, Xianzhi Fu, Rusheng Yuan
PII:
S0021-9517(20)30265-7
https://doi.org/10.1016/j.jcat.2020.06.032
YJCAT 13803
DOI:
Reference:
To appear in:
Journal of Catalysis
Received Date:
Revised Date:
Accepted Date:
24 February 2020
8 June 2020
28 June 2020
Please cite this article as: L. Wang, M. Liu, W. Zha, Y. Wei, X. Ma, C. Xu, C. Lu, N. Qin, L. Gao, W. Qiu, R. Sa,
X. Fu, R. Yuan, Mechanistic study of visible light-driven CdS or g-C3N4-catalyzed C-H direct
trifluoromethylation of (hetero)arenes using CF3SO2Na as the trifluoromethyl source, Journal of Catalysis
(2020), doi: https://doi.org/10.1016/j.jcat.2020.06.032
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Inc.

Mechanistic study of visible light-driven CdS or g-C3N4-catalyzed C-H
direct trifluoromethylation of (hetero)arenes using CF3SO2Na as the
trifluoromethyl source
a
a
a
a
Lele Wang, Ming Liu, Wenying Zha, a,b Yingcong Wei, a Xiongfeng Ma, Chunwang Xu, Chenggang
Lu, a Nanfang Qin, a Li Gao, c Wenzhao Qiu, a Rongjian Sa, *b Xianzhi Fu a and Rusheng Yuan *a
a State Key Laboratory of Photocatalysis on Energy and Environment, Fuzhou University, Fuzhou 350108,
Fujian, China
b Institute of Oceanography, Ocean College, Minjiang University, Fuzhou, Fujian 350108, PR China
c Department of Pharmacy, University of Groningen, Deusinglaan 1, 9700 AV Groningen, Netherlands
*Corresponding author. E-mail: yuanrs@fzu.edu.cn (R. Yuan), rjsa@mju.edu.cn (R. Sa)

Abstract
The mild and sustainable methods for C-H direct trifluoromethylation of (hetero)arenes without any base or
strong oxidants are in extremely high demand. Here, we report that the photo-generated electron-hole pairs
of classical semiconductors (CdS or g-C3N4) under visible light excitation are effective to drive C-H
trifluoromethylation of (hetero)arenes with stable and inexpensive CF3SO2Na as the trifluoromethyl (TFM)
source via radical pathway. Either CdS or g-C3N4 propagated reaction can efficiently transform CF3SO2Na
to •CF3 radical and further afford the desired benzotrifluoride derivatives in moderate to good yields. After
visible light initiated photocatalytic process, the key elements (such as F, S and C) derived from the starting
TFM source of CF3SO2Na exhibited differential chemical forms as compared to those in other oxidative
reactions. The photogenerated electron was trapped by chemisorbed O2 on photocatalysts to form superoxide
-

radical anion (O2 ) which will further attack •CF3 radical with the generation of inorganic product F and
CO2. This resulted in a low utilization efficiency of •CF3 (<50%). When nitro aromatic compounds and
CF3SO2Na served as the starting materials in inert atmosphere, the photoexcited electrons can be directed to
reduce the nitro group to amino group rather than being trapped by O2. Meanwhile, the photogenerated holes
-
oxidize SO2CF3 into •CF3. Both the photogenerated electrons and holes were engaged in reductive and
oxidative paths, respectively. The desired product, trifluoromethylated aniline, was obtained successfully via
one-pot free-radical synthesis.
Keywords: Trifluoromethylation; Semiconductor; (Hetero)Arenes; Free radical; Visible light

Main Text
1. Introduction
Trifluoromethyl (TFM) compounds are core building blocks in the fields of pharmaceutical, organic
synthesis, and agrochemical, since the TFM group can dramatically influence their properties (such as
membrane permeability and bioavailability, elevated electronegativity and oxidation resistance) [1-6]. Thus,
it has gained a noteworthy rise of interest for the incorporation of a -CF3 moiety into organic molecules [7].
Among the diverse synthetic methodologies to construct C-C bonds [8-9], photocatalytic C-H bond direct
trifluoromethylation of (hetero)arenes has been one of the most sustainable and mildest ways as it avoids the
involvement of pre-functionalized substrates and harsh reaction conditions [10-17]. In recent years, effective
visible-light catalysts, with outstanding photocatalytic performance, including homogeneous metal
complexes [18] ([Ru(bpy)3Cl2] [19], [Co(Ph-tpy)2][(PF6)3] [20], Ir(ppy)3 [21]), organic dyes (methylene blue
[22], Eosin Y [23]), quinones [24] and heterogeneous inorganic materials (such as Rh-modified TiO2 NPs)
[25] have been developed in the trifluoromethylation reaction. Semiconductor photocatalysts with diverse
band structures endowed themselves with great potential for driving this reaction [7], [26-27]. However, not
much attention is paid to the trifluoromethylation of (hetero)arenes on cheap and classical semiconductors
without any metal modification or additives (strong oxidants and precious metals).
The radical trifluoromethylation of (hetero)arenes is an efficient strategy owing to the electrophilic
nature of CF3 radical with a low-lying singly occupied molecular orbital (SOMO) [28] which was apt to
react with electron-rich (hetero)arenes. Compared with TFM metals (M=Cu, Zn, Si, Cd, Sn, Hg and Pb)
[29-30] and gaseous CF3I [31], CF3SO2Na as a safe, inexpensive and stable solid TFM source can be easily
oxidized to release CF3 (the lowest potential is 0.6 eV) [24]. Theoretically, visible light responsive
photocatalysts absorb incident photons of the valid light wavelength larger than 420 nm which decided by
their physical bandgaps, and thus generate (oxidative potentials) exciton transitions in the valence band
(VB) being more positive than 0.6 eV enabling the transformation of CF3SO2Na to CF3. Since the
photo-generated excitons is ever in the native form of electron-hole pairs, the separation of these
electron-hole pairs to offer more holes for subsequent oxidation step or efficiently utilize electrons

synchronously is the most important issue in this light-initiated synthetic reactions [32,33]. In addition, the
transfer balance of photo-generated electron-hole pairs and the chemical evolution of the elements in
CF3SO2Na are essential during photocatalytic trifluoromethylation, but little attention has been paid to it
previously.
Herein, we report the application of semiconductors (such as CdS nanorod or g-C3N4) as the
heterogeneous photocatalysts for the functionalization of (hetero)arenes to synthetize trifluoromethylated
(hetero)arenes under visible light irradiation. In this work, a series of benzotrifluoride derivatives were
successfully synthesized in good yields, and the experimental fact revealed that the formed O2 during the
photocatalysis is the key species leading to the unexpected decomposition of CF3 derived from CF3SO2Na.
By introducing nitro aromatic compounds as the starting material, the photo-generated electrons tend to
reduce nitro group of the compounds to amino group. This would avoid the trapping process of electrons by
O2 and be an effective strategy for one-pot synthesis of trifluoromethylated aniline from nitro aromatic
compound.
2. Experimental procedures
2.1. General information
All reagents and solvents are commercially available (Beijing innoChem Science&Technology Co. Ltd.),
and they were further dried by the vacuum rotary evaporator to remove all traces of water and then stored in
the N2-filled glovebox. The 10 mL Schlenk tube charged with a magnetic stirrer, which was used as the
reactor in the trifluoromethylation reaction. The visible light (λ>420 nm) was provided by a 300 W xenon
lamp (PLS-SXE300C, Beijing Perfect Light Co.) equipped with an IR-cutoff filter (λ<780 nm). UV source
was provided by 4-W UV lamp with certain wavelength centered at 365 nm (Philips, TUV 4W/G4 T5). The
LED lights were provided by Beijing Perfect Light Co. (PL-LED100 lamps, centered at 400 ± 15 nm, 420 ±
15 nm, 450 ± 15 nm, 475 ± 15 nm, 500 ± 15 nm, 520 ± 15 nm, 550 ± 15 nm, respectively). The incident
light energy was measured with a spectroradiometer (International Light Technologies Model ILT950).
The chemical structures of the products were confirmed by comparison with standard chemicals and
GC-MS (Agilent Technologies, GC 7890B, MS 5977A) data. GC equipped with a FID detector (Agilent

Technologies, GC 7890 B) and a HP-5 5% phenyl methyl siloxane column (30m  0.32mm  0.5 휇m),
1
13
which was used for the quantifiable measure of trifluoromethoxybenzene. H and C NMR spectra were
recorded on a Bruker UltraShield Plus Avance III (500 MHz and 125 MHz, respectively). Peaks were
referenced to residual solvent (CDCl3, 1H, 7.26 ppm, 13C, 77.0 ppm). The quantitative yields of
trifluoromethylated compounds were calculated using 19F NMR with trifluoromethoxybenzene as the
internal standard. The 19F NMR spectra were obtained at 293 K on a Bruker Avance 500 spectrometer (400
MHz), and chemical shifts were recorded relatively to the solvent resonance.
The size and morphology of catalysts were measured by a Hitachi New Generation SU8010 field
emission scanning electron microscope (FE-SEM, JSM-670 F). The transmission electron microscopy
(TEM) images, high-resolution transmission electron microscopy (HRTEM) images (with an acceleration
voltage of 200 kV), and selected area fast Fourier transform (FFT) pattern were measured on a JEOL
JEM-3000F microscope, which with a CCD camera as the detector. The Brunauer-Emmett-Teller (BET)
surface area and the pore size distribution curve of catalysts were analyzed by a Micromeritics ASAP 2020
surface area analyzer. The X-ray photoelectron spectra (XPS) were acquired using a Thermo Scientific
EscaLab 250 XPS system with a monochromatic Al Kα source and a charge neutralizer. The background
pressure was maintained at <10−6 Pa during data acquisition. The binding energies were referenced to the
hydrocarbon C1s energy of 284.8 eV from adventitious carbon. A Varian Cary 500 UV-Vis
spectrophotometer was used to record the UV-Vis diffuse reflectance spectra (DRS) of various solid
samples. The photoelectrochemical characterization was performed on a Metrohm-Autolab AUT302N
Electrochemical workstation. Photocurrent measurements were carried out in a typical three-electrode
configuration with an Ag/AgCl electrode, a coiled Pt wire as the reference and counter electrode,
respectively. X-ray diffraction (XRD) studies of catalyst were carried out with a Bruker D8 Advance
instrument. The electron paramagnetic resonance (EPR) experiments were carried out on Bruker A300
instrument operating in the X-band at room temperature. An ion chromatograph (Dionex ICS-1100, Dionex)
equipped with a conductivity detector and an ion exchange column (IonPac AS11-HC, Dionex) was used to
detect and quantify F-.
2.2. Catalyst preparation

Synthesis of CdS nanorods. In a typical preparation process, CdCl22.5H2O (4.62 g, 0.023 mol), thiourea
(CH4N2S, 4.62 g, 0.061 mol) and a given amount of ethylenediamine (60 mL) were added into a 100 mL
Teflon-lined stainless steel autoclave. The solution was stirred at room temperature for 1 h, and then heated
to 160 °C for 24 h. When it cooled to room temperature naturally, the product was washed with distilled
water and absolute ethanol several times to remove the excess reactants and byproduct. Finally, the sample
was dried in a vacuum oven at 60 °C for 8 h.
Synthesis of g-C3N4. Typically, it was synthesized thermally by heating urea (10ꢀg) at 550ꢀ°C for 2ꢀh with
a heating rate of 2 °C min-1 under ambient pressure in air. Then, the as-obtained yellow powder was
collected and grinded to get the final sample.
Synthesis of CdS/g-C3N4. Pre-weighed CdS and g-C3N4 were added into the appropriate ethanol-water
mixture (Volume ratio, 1:1). After ultrasonic treatment for 2 h, the mixture was stirred and dried at 60 °C in
an oil bath. Then the obtained powders were calcined at 300 °C in air with a heating rate of 15 °C/min for 2
h. Finally, the bright yellow product was collected for further characterization. The synthesized CdS/g-C3N4
composites are denoted as CS-x, and x represents the theoretical concentration of g-C3N4 (x wt.%). For
example, the 20 wt.% g-C3N4/80 wt.% CdS catalyst can be denoted as CS-80 (Table S1).
2.3. General procedures for the photocatalytic trifluoromethylation of (hetero)arenes
In a typical photocatalytic experiment, benzene (11.2 mmol), CH3CN (1 mL), photocatalyst (10 mg) and
CF3SO2Na (0.5 mmol) were mixed in the reactor in a N2-filled glovebox. The mixture was stirred upon 500
nm LED light irradiation for 12 hours. Then the resulting mixture was centrifuged and the supernatant was
preliminarily analyzed by GC, GC-MS and 19F NMR techniques. All the other photoinduced
trifluoromethylation reactions of (hetero)arenes undergo similar procedures as mentioned above.
2.4. Typical procedure for the thermal reactions
The procedures for thermal reactions of (hetero)arenes and CF3SO2Na are similar to those for the
o
photoreactions while the reaction temperature was controlled by a heating jacket at 60 C under dark
conditions.
2.5. Detection of free radicals
Electron paramagnetic resonance spectroscopy (EPR) was used to monitor the species of free radicals with a


Bruker A300 instrument. Typically, for the detection of O2 , a solution containing CdS (0.07 mmol),
CH3CN (1 mL) and DMPO (5, 5-dimethyl-1-pyrroline N-oxide, 0.1 mmol) was sucked by a capillary tube.
One end of it was plugged with plasticine and the capillary tube was then placed into a sealed quartz tube
under N2 atmosphere. The quartz tube was irradiated with visible light irradiation and the spectra were
recorded in situ.

The detection for CF3 experienced the same procedures as that of O2 , except that MNP
(2-methyl-2-nitrosopropane dimer, 0.1 mmol/mL) was used as the radical scavenger. The sample was
prepared by mixing CF3SO2Na (0.2 mmol/mL) with acetonitrile (1 mL).
2.6. Procedure of KIE experiment between benzene and C6D6
A 10 mL Schlenk tube was charged with benzene (5.6 mmol), C6D6 (5.6 mmol), CH3CN (1 mL) and
CF3SO2Na (0.5 mmol). The resulting mixture was stirred upon visible light (420-780 nm) irradiation for 12
h. The crude mixture was then centrifuged and the supernatant was preliminarily recorded on 19F NMR.
2.7. Computational Details
The proposed reaction path for trifluoromethylation reaction catalyzed by g-C3N4 has been studied by
Density Functional Theoretical (DFT) methods, which were implemented in DMol3 package [34-35]. The
generalized gradient approximation (GGA) with the functional of Perdew-Burke-Ernzerhof (PBE) was
utilized for all calculations. The double-numeric quality basis set to expand wave functions with polarization
functions (DNP) was adopted. The Grimme’s scheme (D2) to treat the long- range dispersion interactions
[36]. A Fermi smearing of 0.005 hartree was utilized in this calculation, and a size of 2 ×2×1 was used for
structural relaxation. The convergence criteria for energy, gradient and displacement convergence were 1.0
× 10−5 Ha, 0.002 Ha/Å, and 0.005 Å, respectively. All periodic slabs had a vacuum spacing of at least 15 Å,
which was thick enough to avoid artificial interaction. A Γ-centered 4×4×3 Monkhorst-Pack grid for the
Brillouin zone sampling was used in our calculations. The adsorption energies (ΔEads) with
zero-point-corrected are defined using the equation Δ퐸ads = [퐸slab+adsorbate − (퐸slab + 퐸adsorbate)] + ΔZPEads,
where Eslab+Eadsorbate, Eslab and Eadsorbate represent the total energy of the slab covered with the adsorbate, the
energy of the bare slab and the energy of the free adsorbate molecules, respectively. ΔZPEads represents the

zero-point vibrational energy (ZPE) correction for the adsorption.
3. Results and discussion
3.1. Photoinduced Trifluoromethylation of (Hetero)Arenes
Initially, a series of semiconductors (such as CdS, g-C3N4, BiVO4, Ag3PO4 and WO3) possessing high
visible light absorption ability (>420 nm) were chosen as the photocatalysts to investigate their
photocatalytic performance in the trifluoromethylation reaction of benzene (Fig. 1a). The corresponding
band structures of these semiconductors showed that the VB positions are more positive than 0.6 eV, which
theoretically enable these semiconductors to be a possible contributor in affording CF3 radicals from
CF3SO2Na (Fig. 1b) [37]. However, when the reaction mixture (2-methyl-2-nitrosopropane (MNP) as a
trapping agent) was irradiated with visible light (420-780 nm, Fig. 1c), only CdS and g-C3N4 could give a
singlet-triplet splitting signal at g=2.0061 (N= F=12.25 G), which can be attributed to MNP-CF3 adduct
[23]. Furthermore, the intensity of this signal increased with a prolonged irradiation time in the CdS
catalyzed system, confirming the generation of CF3 via the oxidation of CF3SO2Na. The
trifluoromethylation reaction proceeded smoothly only in the presence of CdS or g-C3N4, affording 49.2% or
36.8% yield of benzotrifluoride (3a) under visible light irradiation (Table 1, entries 1-6). When the
scavengers of holes (Na2SO3) were added into the CdS reaction system, the trifluoromethylation reaction
was almost terminated. This means that photogenerated holes are solely responsible for the
trifluoromethylation (Table entry 7). These above results suggest that the VB positions and band gaps were
not the sole factors determining the effectiveness of semiconductors in the trifluoromethylation reaction.
To further investigate the reason, we conducted the trifluoromethylation reactions employing CdS as the
model catalyst. As shown in Table 1, no 3a product was observed in the absence of photocatalyst or light,
even upon heating to 60 °C, suggesting that a photocatalytic process was involved in this reaction (entries
8-10). The transformation process could not occur in the absence of O2, indicating that O2 may play an
important role in the occurrence of this reaction (entry 10). Generally, when O2 as the essential substance in

the reaction system, it could be reduced to superoxide radical anion (O2 ) by the semiconductors which has
negative enough CB positions ( -0.33 eV) under light irradiation (Fig. 1b). Thus, we reason that, in the case

of O2 the photo-generated electrons can be trapped by O2 when the semiconductors possess negative enough

CB positions than -0.33 eV [38]. This speculation was then testified by the EPR spectra of O2 , in which

characteristic signals for DMPO-O2 adducts were observed when acetonitrile solution containing CdS or
g-C3N4 (with 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the trapping agent) was irradiated by visible
light (Fig. 1d). Similarly, the signal intensity increases with the irradiation time. No O2 signal was obtained
in the presence of BiVO4, Ag3PO4 or WO3 [39]. This trend was in good agreement with the experimental
result of these semiconductors in the trifluoromethylation reaction. Thus, we rationally concluded that O2 as
an electron trapping agent prevented the fast recombination of electrons and holes, benefitting the formation
of 3a.
Fig. 1. UV-vis DRS spectra (a) and the band structures (b) of semiconductors. ESR spectra of trifluoromethyl radicals (c)
and super oxide anion (d).

Table 1 Photocatalytic C-H direct trifluoromethylation of benzene.
Entry
Condition
Photocatalyst Amount of 3a (mol)
Yield (%)
1a
2
420-780 nm
420-780 nm
420-780 nm
CdS
246
184

49.2
36.8

g-C3N4
BiVO4
3
4
5
420-780 nm
420-780 nm
420-780 nm
420-780 nm
Dark
Ag3PO4
WO3
none
CdS

trace


trace

6
7b
8
trace

trace

CdS
9
60 oC
CdS


10c
11d
12e
420-780 nm
500 nm
CdS
CdS
CdS


210
118
42.0
23.6
Household light
a
Reaction conditions: 1a (11.2 mmol, 22.4 equiv), 2a (0.5 mmol, 1 equiv), CH3CN (1 mL), photocatalyst (10 mg), room
temperature (25 °C), air atmosphere (0.1 MPa), irradiation time (24 h), GC yields (Yield = the molar ratio of benzotrifluoride
to initial CF3SO2Na), “” means not detected. Light intensity (500 mW/cm2). b Na2SO3 is added in the reaction system.
c N2 atmosphere. d Light intensity (80 mW/cm2). e Light intensity (20 mW/cm2, λ= 380-700 nm).
Inspired by these initially promising results, we further explored the effects of different variables on the
activity of CdS. Upon irradiation with a 500 nm LED light with a lower intensity, nearly no significant
decrease in the yield of 3a (42%) was observed (Table 1, entry 11). Even household CFL lamp could
promote the reaction and gave a moderate yield of 23.6% (entry 12). As a substitute for CH3CN, other
solvents such as CH3OH, EtOH, acetone, DMSO and THF could also promote the reaction in moderate
yields which were slightly lower than that of CH3CN (Fig. 2a). Surprisingly, the introduction of base or acid
such as K2HPO4, K2CO3, CH3COOH, CF3COOH into the reaction resulted in a decrease in the yield of 3a,

suggesting that the additives were not needed for a higher yield of benzotrifluoride (Fig. 2b). As shown in
Fig. 2c, the amount of 3a linearly increased with both the increasing light intensity and the prolonged
irradiation time. The dependence of apparent quantum yields (AQYs) on the wavelength of excitation light
in this reaction was further investigated. As shown in Fig. 2d, the trend of AQYs at different wavelengths
matched well with that of the absorption spectra. The highest AQY (0.43%) of CdS was obtained at 500 nm
within the visible light range. The longest wavelength available for the reaction has been extended to 550
nm which corresponds to the absorption edge of the catalyst. When 600 nm LED light was used, no product
was obtained. The effective wavelength range is compatible with the absorption spectrum of CdS, further
indicating that the photo-generated holes over CdS are responsible for the trifluoromethylation of benzene.
Fig. 2. (a) Yields of 3a in different solvents. (b) The influence of base or acid (0.25 mmol) on the amount of 3a irradiated
for 14 h. (c) The influence of irradiation time and light intensity (inset, illumined for 16 h) on the amount of 3a. (d) AQYs
at different single-wavelengths and the UV-Vis spectrum of CdS.
The scope of this trifluoromethylation reaction was successfully extended to a wide variety of
(hetero)arenes (Table 2). Most of the employed aromatic substrates could gave moderate yields of target

products. When benzene ring bearing electron-donating substituents (such as -OMe, -Me, -SMe and -Et)
gave higher yields than that bearing the electron-withdrawing ones (-CN, -I, -Br, 3b-h), and the yields of
desired products increased with the enhancement of donor power (or decreased electro-withdrawing power)
of substituent groups. This also confirmed that the trifluoromethylation reaction is sensitive to the electronic
effect of the substituent group. Interestingly, the liable leaving group such as -I can keep stable during the
trifluoromethylation process since no benzotrifluoride was detected when using iodobenzene as the substrate
(3h). The trifluoromethylation of disubstituted benzene (3b, 3i, 3j, 3c, 3k) could also take place and give
12.4-39.4% yield of desired products, respectively. Additionally, a wide range of ring substitution and
heteroatom protecting groups are well-tolerated affording 7-33.4% yields of target products (3l-3q).
Five-membered and six-membered heterocycles including benzofuran, benzimidazole, benzothiophene
could also be activated and afford the desired trifluoromethylated products which can be potential
pharmacophores. These results suggest that, this heterogeneous photocatalytic system possess good
tolerance of functional groups.
Table 2 Reaction scope of C-H direct trifluoromethylation of (hetero)arenes

Reaction conditions: 1 (0.2 mL), 2a (0.5 mmol), CH3CN (1mL), photocatalyst (10 mg), air atmosphere, light intensity (80
mW/cm2, λ=500 nm), irradiation time (24 h), room temperature (25 °C), air atmosphere (0.1 MPa) 19F NMR yields with
trifluoromethoxybenzene as internal standard.
3.2. Photocatalyst Characterization
The more detailed characterization of the morphologies and microstructures of the photocatalysts were
obtained by SEM and TEM measurements. As shown in Fig. 3a, b, rod-like CdS with an average diameter of
40 nm and the length in the range of 200-300 nm was obtained. The high-resolution TEM images (Fig. 3c)
in the selected regions of CdS nanorod showed that the sample exhibits good crystalline and clear lattice
fringes with an interval of 3.36 Å corresponding to the interatomic distance of hexagonal CdS (002) [40,41].
The crystal orientation of the single crystalline CdS was also confirmed by the fast Fourier transform (FFT)
pattern. The TEM and SEM images of g-C3N4 showed that the obtained g-C3N4 samples are aggregated by
irregular nanosheets (Fig. 3d-f).
Fig. 3. (a-c) The SEM, TEM and HRTEM images of CdS sample. Inset of (b) is the selected area Fast Fourier transform
(FFT) pattern. (d-f) The SEM and TEM images of g-C3N4 sample.
The surface chemical composition and oxidation states of the CdS and g-C3N4 samples were further
monitored by XPS measurement. As shown in Fig. 4a, the survey spectrum of the elemental composition for
CdS confirmed that it was composed of Cd and S elements. The peaks at 404 eV and 410.8 eV in the
high-resolution XPS spectrum of the Cd element can be assigned to Cd 3d3/2 and Cd 3d5/2, respectively (Fig.
4b), [42]. The 6.7 eV difference between these two peaks confirmed the Cd2+ states. The peaks at 160.5 eV

and 161.5 eV in Fig. 4c correspond to S 2p transitions. The g-C3N4 is comprised of C and N elements as
evidenced in the survey spectrum (Fig. 4d). The high-resolution spectrum of the C element showed two
peaks at about 284.6 eV and 287.9 eV, which were ascribed to the C in C-C and the sp2-bonded C in C=N,
respectively (Fig. 4e) [43]. As shown in Fig. 4f, the N1s spectrum could be deconvoluted into three peaks
centered at 398.4, 400 and 400.5 eV, respectively. The peak at 398.4 eV was assigned to the N in the sp2
C=N bond and the peaks at 400 and 400.9 eV were corresponding to the N atoms bounded to three C atoms.
These results were similar to that of the literature [44]. Other physicochemical properties such as pore
characteristics and the specific surface area of CdS and g-C3N4 materials were also studied by N2
physisorption measurements (Fig. 4). The BET specific surface area of CdS and g-C3N4 were calculated to
be 14.2 m²/g and 94.1 m²/g, respectively. This also indicated that the difference in physical adsorptivity of
CdS and g-C3N4 was not the major reason resulting in the better performance for CdS in
trifluoromethylation of arenes. Further photocurrent spectrum of CdS exhibited a higher response compared
to that of g-C3N4, suggesting better separation ability of the electron-hole pairs (Fig. 5). Considering the
preferable photocatalytic performance of CdS under the same reaction conditions, one can conclude that the
improved activity of CdS profits from the enhanced visible light absorption and efficient inhibition of the
charge carrier recombination. Generally, the separation efficiency of the photoexcited electron-hole pairs
will be significantly improved by creating a p-n heterojunction at the interface of p- and n-type
semiconductors. This stimulated us to further prepare CdS/g-C3N4 composite photocatalyst with joint
interface between CdS nanorods and g-C3N4 sheets under optimized synthetic condition as shown in Fig. 6a
and 6b (Fig. S1, S2). As expected, the composite catalysts gave significant improvement in the conversion
rate of benzotrifluoride as compared to that of pure CdS or g-C3N4 sample (Fig. 6c and 6d). Among all the
CdS/g-C3N4 composites, CS-60 showed the highest rate. However, the final yield of the trifluoromethylated
products with prolonged reaction time for all the composites were still lower than 50% of CF3SO2Na (Table
S2).

Fig. 4. (a, d) XPS survey spectra of CdS and g-C3N4. (b, c, e, f) High resolution XPS spectra of Cd 3d, S 2p, C 1s and N 1s
in CdS and g-C3N4 samples, respectively.
Fig. 5. Photocurrent spectra of CdS (a) and g-C3N4 (b) samples.

Fig. 6. TEM (a), HRTEM (b) image for CS-60. (c) Yields of 3a using different catalysts for 8 h. (d) The influence of
irradiation time on the amount of 3a.
3.3. Cyclic Utilization of the Photocatalyst
To investigate the stability of CdS and g-C3N4 photocatalysts, 5 cycles of photocatalyzed
trifluoromethylation were conducted. As shown in Fig. 7a, there is almost no obvious decrease in the yield
of 3a after 5 reaction cycles using g-C3N4 as the photocatalyst, and only a slight decrease can be found for
CdS. The TON value for CdS was up to 6.8. Interestingly, g-C3N4 contribute to a rather stable photocatalytic
performance towards the trifluoromethylation of (hetero)arenes. The crystal structures of these
photocatalysts before and after 5 reaction cycles were further revealed by XRD patterns. It is obvious that
the diffraction peaks of fresh CdS and g-C3N4 can be assigned to hexagonal CdS phase and graphite phase
by comparing with the standard data from JCPDS card (File No. 41-1049 and 87-1526), respectively (Fig.
7b, c) [45,46]. All peaks appeared in the recycled CdS and g-C3N4 generally match the standard phases of
fresh CdS and g-C3N4, suggesting that the crystallization and crystallite size of these catalysts are not
changed by visible light irradiation.

Fig. 7. (a) The recycle experiments of CdS and g-C3N4. (b, c) XRD patterns of fresh CdS and recycled CdS for five cycles.
3.4. Reaction Mechanism Investigation
Addition of tetramethylpiperidine N-oxide (TEMPO, 2 equiv), a typical radical scavenger, to the standard
reaction system significantly lower the yield of 3a (Table 3). This coupling with the evidence of achieving
MNP-CF3• adduct in radical trapping experiment leads to the confirmation of a radical dominant pathway
for the trifluoromethylation reactions of (hetero)arenes over CdS or g-C3N4 photocatalyst. The radical
involving steps focusing on •CF3 over solid g-C3N4 surface was also studied by DFT calculations. The top
and side views of g-C3N4 structures and the corresponding complex structures with various species (C6H6,
CF3, C7H5F3, H) involved in the trifluoromethylation reaction adsorbed on g-C3N4 are optimized (Fig. S3).
The calculated free energy diagrams of trifluoromethylation reaction through the above reaction path on
g-C3N4 is summarized in Fig. 8. The negative Eads indicates that the adsorption is an exothermic process in
all instances. The Eads values of adsorbates are all larger than 1 eV, which indicates the strong
chemisorption of these species on g-C3N4. The thermodynamic profile shows that the formation of C7H5F3
along the step (C6H6*+CF3*
C7H5F3*+H*) undergoes an energy release process. Additionally, a
competitive experiment with the same amounts of benzene and C6D6 (1a-d6) in the presence of CdS or
g-C3N4 was then carried out under visible light irradiation. The kinetic isotopic effect of benzene showed a
low KH/KD value of 1.0 for the transformation, suggesting that the cleavage of C-H bond in benzene was not
the rate-determining step (Fig. 9).
Table 3 Radical trapping experiments

Entry
Addictive
none
Amount of 3a (μmol)
1
2
109
<1
TEMPO (2 equiv)
Reaction conditions: 1a (11.2 mmol), 2a (0.5 mmol), CH3CN (1 mL), photocatalyst (20 mg), air atmosphere, light intensity (80
mW/cm2, λ=500 nm), irradiation time (12 h), room temperature (25 °C), air atmosphere (0.1 MPa), TEMPO
=2,2,6,6-tetramethylpiperidine-N-oxyl, GC yields.
Fig. 8 The Eads of species involved in trifluoromethylation of benzene with g-C3N4, inset was the optimized structures of
species adsorbed on the monolayer g-C3N4.

Fig. 9. KIE experiment.
Based on the above experimental and DFT results, we then shifted our attention to the activation
mechanism of CF3SO2Na under O2 atmosphere during the trifluoromethylation reaction. After light
irradiation, F- and 3a were detected in the ion chromatography spectrum and the 19F NMR spectrum,
respectively. The amounts of F- and 3a increased along with the extending irradiation time with a
growth rate ratio of 3:1, indicating that F- and benzotrifluoride were the final chemical form of F elements in
CF3SO2Na after the transformation (Fig. 10a). When the gas phase reaction mixture was measured on
GC-MS, SO2 and CO2 were detected while none of them were found in the absence of CF3SO2Na (Fig. S4).
2-
2-
No SO4 was detected by the ion chromatography, thus the formation of SO4 was ruled out. This was
different from the result we have reported using photoexcited quinones as the catalyst in the photocatalytic
2-
trifluoromethylation of (hetero)arenes, in which CF3SO2Na transformed into final inorganic product S2O4 ,
followed by a classical hydrolyzation reaction to give SO42- in the presence of H2O and O2 [24]. According
to the above evidences, the whole chemical equation was given in Fig. 10b. This process catalyzed by
semiconductor having a strong reductive ability exhibited a lower efficiency in F element utilization than
quinones with strong oxidation.

Fig. 10. Influence of irradiation time on the amount of benzotrifluoride and fluorine anion (a). The chemical equation of the
trifluoromethylation reaction (b).
On the basis of all the aforementioned mechanistic investigation, a plausible mechanism for photo-driven
semiconductor catalyzed trifluoromethylation of (hetero)arenes using CF3SO2Na as the CF3 source is
proposed (Scheme 1a). Upon excitation, the photo-generated hole in the valence band of semiconductor

oxidized the CF3SO2 into CF3 and SO2, which is the initiation step of the reaction. Then the CF3 followed
two reaction pathways as shown in Scheme 1. The electron-deficient CF3 attacks on the electron-rich
position of (hetero)arenes to produce the key intermediate of trifluoromethyl-cyclohexadienyl radical via
electrophilic aromatic substitution (SEAr). In this process, more than one electron-rich site in the substrates
coupled with the steric hindrance effect of the substituent group resulted in low regioselectivity of some
reactions. Finally, the formed intermediate was oxidized by the separated hole, followed by the

deprotonization process to give the desired benzotrifluoride. Alternatively, the CF3 reacted with O2 and
electrons giving F- and CO2, which result in a low utilization efficiency of CF3SO2Na. This photocatalytic
strategy tolerates oxidative reaction mode using O2 as the electron acceptor [25]. During this process, the

photo-generated electrons made no direct contribution to the activation of the substrate or participation in
yielding the desired product. One can expect that, once a suitable reductive half-reaction process was
introduced into this system, a full utilization of the photo-generated electron-hole pairs rather than using the
holes only will be realized. Thus, a trifluoromethylation reaction using nitroaromatic compound as the
substrate to synthesize trifluoromethylated anilines was designed to testify this rational speculation (Scheme
1b).
Scheme 1 Proposed mechanism of the trifluoromethylation reaction.
Trifluoromethylated anilines as the core pharmaceutical intermediates have attracted many interests [47].
Current strategies for the synthesis of them mainly use amines as the substrate in the presence of strong
acids or bases to protect the amino group which makes the transformation more complicated. We proposed a
one-pot synthesis of trifluoromethylated anilines from nitro aromatic compound under mild reaction
conditions without functional group protection process. When using the trifluoromethylation of
4-chloronitrobenzene (4a) as the model reaction, CdS and g-C3N4 gave 23% and 10% yield of
2-trifluoromethyl aniline (5a) expectively in the presence of CH3OH and formic acid under visible light
irradiation for 24 hours (Table 4, entries 1-2). As an initial attempt, the transformation hardly proceeded in
the absence of CH3OH or formic acid (Table 4, entry 3), which indicates the complicity of this multiphase
reaction. Nitro compound was reduced to aromatic amine [48,49] by photo-generated electrons, avoiding the

decomposition reaction of •CF3 radical by O2 . As shown in Fig. S5, 4-Chloroaniline as main reductive
by-product was found with a yield of 59 mol determined by GC analysis. However, non-reduction product

was hardly detected. It is probably because -NO2 was easy to reduce as -NH2 in this reaction system.
Table 4 One-pot synthesis of trifluoromethylated anilines from nitro aromatic compounda
Entry
Substrate
Photocatalyst
CdS
Product
Amount (μmol) Yield (%)
1a
2
4a
4a
4a
5a
5a
5a
115 c
50
23
10
--
g-C3N4
CdS
3b
--
a Reaction conditions: 4a (0.2 mmol), 2a (0.5 mmol), CH3CN (1mL), MeOH (1.5 mmol), formic acid (1.5 mmol), photocatalyst
(10 mg), light intensity (500 mw/cm2, λ=420-780 nm), irradiation time (24 h), room temperature (25 °C), N2 atmosphere (0.1
b
c
MPa). 19F NMR yields with trifluoromethoxybenzene as internal standard. without MeOH or formic acid. 59 mol
p-chloroaniline as by-product also detected.
4. Conclusions
We have developed a reliable route for direct C-H trifluoromethylation reaction of (hetero)arenes with a
broad substrate scope through CdS or g-C3N4 initiated photocatalysis by using CF3SO2Na to provide •CF3
radicals. The experimental results revealed the unwanted side reaction of O2  with •CF3 to form CO2 and F-,

which limited the effective utilization of CF3SO2Na in photocatalytic synthesis. Furthermore, one-pot
synthesis of trifluoromethylated anilines using nitro aromatic compound and CF3SO2Na as the
multifunctional substrates has been provided to utilize photogenerated electron-hole pairs simultaneously in
one reaction. The understanding of the reaction mechanism of CF3SO2Na in the semiconductor catalyzed
reactions helps us to further design new classes of synthetic organic reactions relative to •CF3 in
photocatalytic mode.
Conflicts of interest
There are no conflicts to declare.
Acknowledgments
This work was financially supported by the National Nature Science Foundation of China 21872033 and

21643009.
References
[1] T. Furuya, A. S. Kamlet and T. Ritter, Nature 473 (2011) 470-477.
[2] C. Le, T. Q. Chen, T. Liang, P. Zhang and D. W. C. MacMillan, Science 360 (2018) 1010-1014.
[3] X. Liu, C. Xu, M. Wang and Q. Liu, Chem. Rev. 115 (2015) 683-730.
[4] E. J. Cho, T. D. Senecal, T. Kinzel, Y. Zhang, D. A. Watson and S. L. Buchwald, Science 328 (2010)
1679-1681.
[5] N. D. Ball, J. W. Kampf and M. S. Sanford, J. Am. Chem. Soc. 132 (2010) 2878-2879.
[6] Y. Ye, S. A. Künzi and M. S. Sanford, Org. Lett. 14 (2012) 4979-4981.
[7] I. Ghosh, J. Khamrai, A. Savateev, N. Shlapakov, M. Antonietti and B. König, Science 365 (2019)
360-366.
[8] M. Hiroki, S. Terajima and T. Shishido, ACS Catal. 8 (2018) 6246-6254.
[9] D. D. Beattie, A. C. Grunwald, T. Perse, L. L. Schafer and J. A. Love, J. Am. Chem. Soc. 140 (2018)
12602-12610.
[10] Y. Ji, T. Brueckl, R. D. Baxter, Y. Fujiwara, I. B. Seiple, S. Su, D. G. Blackmond and P. S. Baran,
PNAS 108 (2011) 14411-14415.
[11] A. Studer, Angew. Chem. Int. Ed. 51 (2012) 8950-8958.
[12] L. L. Zhao, P. H. Li, X. Y. Xie and L. Wang, Org. Chem. Front. 5 (2018) 1689-1697.
[13] L. Li, X. Y. Mu, W. B. Liu, Y. C. Wang, Z. T. Mi and C. J. Li, J. Am. Chem. Soc. 138 (2016)
5809-5812.
[14] D. A. Nagib and D. W. C. MacMillan, Nature 480 (2011) 224-228.
[15] E. Torti, S. Protti and M. Fagnoni, Chem. Commun. 54 (2018) 4144-4147.
[16] P. Liu, W. B. Liu and C. J. Li, J. Am. Chem. Soc. 139 (2017) 14315-14321.
[17] Tong Y, Pan H, Huang W, et al. New J. Chem. 43 (2019) 8741-8745.
[18] O. A. Tomashenko and V. V. Grushin, Chem. Rev. 111 (2011) 4475-4521.
[19] N. J. W. Straathof, H. P. L. Gemoets, X. Wang, J. C. Schouten, V. Hessel and T. Noël, ChemSusChem
7 (2014) 1612-1617.
[20] A. K. Pal, C. F. Li, G. S. Hanan and E. Zysman-Colman, Angew. Chem. Int. Ed. 57 (2018) 1-6.

[21] N. J. W. Straathof, S. E. Cramer, V. Hessel and T. Noёl, Angew. Chem. Int. Ed. 55 (2016) 15549
-15553.
[22] S. P. Pitre, C. D. McTiernan, H. Ismaili and J. C. Scaiano, ACS Catal. 4 (2014) 2530-2535.
[23] N. J. W. Straathof, D. J. G. P. van Osch, A. Schouten, X. Wang, J. C. Schouten, V. Hessel and T. Noël,
J. Flow Chem. 4 (2014) 12-17.
[24] B. Chang, H. G. Shao, P. Yan, W. Z. Qiu, Z. Q. Weng and R. S. Yuan, ACS Sustainable Chem. Eng. 5
(2017) 334-341.
[25] J. Lin, Z. Li, J. Kan, S. J. Huang, W. P. Su and Y. D. Li, Nat. Commun. 8 (2016) 14353.
[26] M. Baar and S. Blechert, Chem. Eur. J. 21 (2015) 526-530.
[27] M. Zhao, H. Xu, H. Chen, S. Ouyang, N. Umezawa, D. Wang and J. Ye, J. Mater. Chem. A 3 (2015)
2331-2337.
[28] J. Tauber, D. Imbri and T. Opatz, Molecules 19 (2014) 16190-16222.
[29] H. Morimoto, T. Tsubogo, N. D. Litvinas and J. F. Hartwig, Angew. Chem. 123 (2011) 3877-3882.
[30] M. A. García-Monforte, S. Martínez-Salvador and B. Menjón, Eur. J. Inorg. Chem. 2012 (2012)
4945-4966.
[31] T. Kino, Y. Nagase, Y. Ohtsuka, K. Yamamoto, D. Uraguchi, K. Tokuhisa and T. Yamakawa, J.
Fluorine Chem. 131 (2010) 98-105.
[32] H. W. Huang, B. Pradhan, J. Hofkens, M. B. J. Roeffaers and J. A. Steele, ACS Energy Lett. 5 (2020)
1107–1123.
[33] A. Dalui, M. Pandey, P. K. Sarkar, B. Pradhan, A. Vasdev, N. B. Manik, G. Sheet and S. Acharya. ACS
Appl. Mater. Interfaces 11 (2019) 11749-11754.
[34] J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865.
[35] A. Bergner, M. Dolg, W. Küchle, H. Stoll and H. Preuß, Mol. Phys. 80 (1993) 1431-1441.
[36]C. -S. Lee, T.-S. Hwang, Y. Wang, S.-M. Peng and C.-S. Hwang, J. Phys. Chem. 100 (1996)
2934-2941.
[37] J. T. Li and N. Q. Wu, Catal. Sci. Technol. 5 (2015) 1360-1384.
[38] S. M. Goodman, M. Levy, F. F. Li, Y. C. Ding, C. M. Courtney, P. P. Chowdhury, A. Erbse, A.
Chatterjee and P. Nagpal, Front. Chem. 6 (2018) 46.

[39] D. Dvoranová, Z. Barbieriková and V. Brezová, Molecules 19 (2014) 17279-17304.
[40] Y. -W. Jun, S.-M. Lee, N. -J. Kang and J. Cheon, J. Am. Chem. Soc. 123 (2001) 5150-5151.
[41] Y. C. Liang and T. W. Lung, Nanoscale Res. Let., 11 (2016) 1-11.
[42] J. Yu, C. Gong, Z. Wu, Y. Wu, W. Xiao, Y. Su, L. Sun, and C. J. Lin, J. Mater. Chem. A 3 (2015)
22218-22226.
[43] W. Zhang, L. Zhou and H. Deng, J. Mater. Chem. A 423 (2016) 270-276.
[44] E. Vorobyeva, Z. Chen, S. Mitchell, R. K. Leary, P. Midgley, J. M. Thomas, R. Hauert, E. Fako, N.
López and J. Pérez-Ramírez, J. Mater. Chem. A 5 (2017) 16393-16403.
[45] H. Q. Cao, G. Z. Wang, S. C. Zhang, X. R. Zhang and D. Rabinovich, Inorg. Chem. 45 (2006)
5103-5108.
[46] J. Ma, C. Wang and H. He, Appl. Catal. B-Environ. 184 (2016) 28-34.
[47] X. Y. Gao, Y. Geng, S. J. Han, A. Liang, J. G. Li, D. P. Zou, Y. S. Wu, and Y. J. Wu, Org. Lett. 20
(2018) 3732-3735.
[48] T. Cuypers, P. Tomkins and D. E. De Vos, Catal. Sci. Technol. 8 (2018) 2519-2523.
[49] S. Li, F. Wang, Y. Liu and Y. Cao, Chin. J. Chem. 35 (2017) 591-595.

TOC
Highlights
• The CdS or g-C3N4 semiconductor photocatalyst could effectively propagate trifluoromethylation reaction
of (hetero)arenes with CF3SO2Na as the trifluoromethyl source under visible light irradiation.

• The formed O2 limited the utilization efficiency of CF3SO2Na in semiconductor catalyzed
trifluoromethylation reaction.
• Introduction of nitro aromatic compounds as the substrate avoided the unwanted side reaction of
photogenerated electrons.
• One-pot synthesis of trifluoromethylated anilines with nitro aromatic compounds as the substrate can been
accomplished employing CdS or g-C3N4 as the photocatalyst.