Continuous process for production of CuCF3 via direct cupration of fluoroform

Article abstract of DOI:10.1021/op500109v

The first continuous flow process has been developed for the synthesis of a superior trifluoromethylating reagent, ligandless CuCF₃, from fluoroform, by far the best CF₃ source in terms of availability, cost, and atom economy. Optimization of the residence time and feed rates for CHF₃ (gas), the cuprating reagent (premade from CuCl and t-BuOK in a 1:2 ratio) in DMF, and the stabilizer (Et₃N· 3HF) at 23 °C and atmospheric pressure has allowed for the continuous production of CuCF₃ in consistently high yields of up to 94%. The thus produced CuCF₃ has been shown to be as highly efficient a trifluoromethylating agent as the one from the previously developed batch process.

Full text of DOI:10.1021/op500109v

Article
pubs.acs.org/OPRD
Continuous Process for Production of CuCF3 via Direct Cupration of
Fluoroform
Zahra Mazloomi, Atul Bansode, Pedro Benavente, Anton Lishchynskyi, Atsushi Urakawa,*
and Vladimir V. Grushin*
Institute of Chemical Research of Catalonia (ICIQ), Tarragona 43007, Spain
ABSTRACT: The first continuous flow process has been developed for the synthesis of a superior trifluoromethylating reagent,
“ligandless” CuCF3, from fluoroform, by far the best CF3 source in terms of availability, cost, and atom economy. Optimization of
the residence time and feed rates for CHF3 (gas), the cuprating reagent (premade from CuCl and t-BuOK in a 1:2 ratio) in
DMF, and the stabilizer (Et3N·3HF) at 23 °C and atmospheric pressure has allowed for the continuous production of CuCF3 in
consistently high yields of up to 94%. The thus produced CuCF3 has been shown to be as highly efficient a trifluoromethylating
agent as the one from the previously developed batch process.
>90% yield, while readily occurring at ambient temperature and
pressure. Second, the cupration process employs only low-cost
materials, the most expensive chemical used being potassium
tert-butoxide. Finally, we have demonstrated the exceptional
efficiency of the fluoroform-derived CuCF3 reagent in a variety
of high-yielding trifluoromethylation reactions of aryl halides,8,9
aryl boronic acids,10 α-haloketones,11 and some other
substrates.8 In our nonindustrial setting, the synthesis of
CuCF3 from fluoroform has been performed in a batch manner
on up to a 0.1 mol scale.9 Considering the potential of
fluoroform-derived CuCF3 and numerous advantages of
continuous over batch operations,12 we explored the possibility
of performing the cupration of fluoroform in a flow manner.
Herein we report the results of our effort that has led to the
development of a reliable continuous process for the
production of CuCF3 from CHF3 in up to 94% yield. We
also demonstrate that the CuCF3 prepared in our flow setup is
as efficient a trifluoromethylating agent as the one made in a
batch manner.
INTRODUCTION
■
Trifluoromethylated organic compounds play an important role
in the synthesis of modern pharmaceuticals, agrochemicals, and
specialty materials.1,2 Over the last 25 years, considerable
progress has been made toward the development of efficient
and selective methods for the introduction of the CF3 group
into a variety of organic substrates.2 The vast majority of these
methods, however, employ cost-prohibitive trifluoromethylat-
ing reagents and therefore cannot be used on an industrial
scale.3 Fluoroform (CHF3, trifluoromethane, HFC-23, R-23) is
by far the best source of the CF3 group for the following
reasons.2r,4
(1) Fluoroform is readily available and cheap. The
fluorochemical and fluoropolymer industries side-
generate large quantities of HFC-23 in chlorodifluoro-
methane (HCFC-22, R-22) manufacture. Approximately
20,000 t of HFC-23 that are produced annually5 lack
applications on a commensurate scale.
(2) Fluoroform is nontoxic and ozone-friendly. Unlike many
polyfluorinated small molecules, CHF3 is not on the
Montreal Protocol list of substances that deplete the
ozone layer.
(3) As a CF3 source for trifluoromethylation reactions,
fluoroform is second to none in terms of atom economy.
(4) The extremely high warming potential of HFC-23
(11,700 times that of CO2) and its long atmospheric
lifetime (>250 years) have raised serious environmental
concerns.5−7 To eliminate the risk of the so-called
“climate bomb”,7 HFC-23 waste streams cannot be
released into the atmosphere and therefore should be
treated. The obviously preferred alternative to the costly
destruction of HFC-23 is its use as a chemical feedstock
for valuable fluorinated compounds.
RESULTS AND DISCUSSION
■
Background and General Considerations. The prepara-
tion of CuCF3 from fluoroform consists of three steps that are
shown in Scheme 1. In Step 1, a solution of the cuprating
reagent is prepared by agitating CuCl and t-BuOK in a 1:2
molar ratio in DMF. The reaction is usually complete within
20−30 min in a batch reactor to give quantitatively the
dialkoxycuprate [K(DMF)][Cu(OBu-t)2] (1), whose solid
state structure has been previously established by single-crystal
diffraction.8 As 1 goes into the DMF solution during the
reaction, the KCl byproduct precipitates out. Although we
conventionally separate the KCl precipitate by filtration, this is
not critical for the next step and the overall CuCF3 synthesis.
What is critical, however, is to maintain rigorously anhydrous,
However, activation of highly inert fluoroform in a
chemoselective manner and with low-cost materials for
industrial applications is a considerable challenge.4 The recently
discovered reaction of direct cupration of fluoroform8 holds
promise as a transformation that might appear suitable for
larger-scale operations. First, this reaction furnishes CuCF3 in
Special Issue: Fluorine Chemistry 14
Received: March 28, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Reaction sequence for the cupration of fluoroform
air-free conditions throughout the preparation and use of 1
since this reagent is highly moisture- and O2-sensitive.
A DMF solution of 1 prepared in Step 1 is then used to
cuprate fluoroform in Step 2. The reaction of CHF3 with 1
occurs smoothly and quickly at room temperature and
atmospheric pressure to give the CuCF3 product within
minutes. If carried out properly, the yield of the cupration
reaction is reproducibly close to quantitative. Although the
product of Step 2 is too unstable for isolation and structural
characterization, sufficient evidence has been obtained13,14 for
this compound being a mixed cuprate bearing one CF3 and one
t-BuO ligand on the Cu(I) center, [K(DMF)n]
[(CF3)Cu(OBu-t)]. In particular, the treatment of the reaction
solution obtained in Step 2 with 18-crown-6 has been shown13
to give stable [K(18-crown-6)][(CF3)Cu(OBu-t)] (X-ray
structure). Furthermore, the closely related cupration reaction
of C2F5H produces [K(DMF)2][(C2F5)Cu(OBu-t)] that has
also been isolated and structurally characterized.14
Figure 1. Flow process setup for the production of CuCF3 from
fluoroform.
Table 1. Cupration of CHF3 in a flow reactor
tube
CHF3,
CHF3 addition
rate, mL min−1
residence
time, min
CuCF
3b
a
entry length, m equiv
yield, %
Once formed in Step 2, the mixed cuprate product begins to
decay8 and should be promptly stabilized with Et3N·3HF
(TREAT HF) or another HF source in Step 3. This
stabilization serves the dual purpose of (i) sequestering, in
the form of KF, the potassium ions that decompose the CF3
group on Cu via α-F-elimination and (ii) acidolyzing the
t-BuO-Cu bond to suppress the unwanted competition of tert-
butoxylation with the desired trifluoromethylation, as pre-
viously observed for some substrates.8,9 The “ligandless” CuCF3
reagent thus prepared can be stored in an inert atmosphere at
room temperature for a few days and at −30 °C for months
without signs of significant decomposition.8
Flow Process Design and Optimization. The above
analysis provided guidelines for the development of a
continuous process to produce CuCF3 from fluoroform. In
the current work, the cupration and stabilization steps (Steps 2
and 3 in Scheme 1) were selected for continuous process
development.15 The cuprating reagent 1 was prepared
separately and fed into the flow reactor using a syringe
pump, as shown in Figure 1. Fluoroform gas was fed into the
reactor through a separate line (Figure 1) by means of a mass
flow controller (MFC). All experiments were carried out at
room temperature (∼23 °C) and atmospheric pressure under
argon. The results of a series of runs toward optimization of
process efficiency and the yield of CuCF3 are compiled in Table
1. In all experiments, the concentration of 1 in DMF (0.48 M)
was close to the solubility limit. The yield of the CuCF3 was
monitored by 19F NMR analysis with an internal standard every
5−10 min during each run. In the first 2−3 min, the yield of
CuCF3 was often lower because of the presence of residual
traces of air in the system and the time required for the addition
and efficient mixing of the reagents in the reactor to
synchronize and level off. During this short initial period
prior to system stabilization, the yield invariably increased to
plateau for the rest of the run. It is these yields of fluoroform-
1
0.61
1.04
1.40
1.52
1.53
1.95
2.32
2.73
1.10
1.32
1.53
2
6
1.25
2.00
2.50
2.75
6.00
8.83
9.33
11.33
5.58
8.83
11.83
85
87
94
94
90
92
94
88
75
86
79
2
2
6
3
2
6
4
2
6
5
1.5
1.5
1.5
1.5
1.2
1.2
1.2
4.5
4.5
4.5
4.5
3.6
3.6
3.6
6
7
8
9
10
11
a
In all experiments, the rates of addition of 1 in DMF (0.48 M) and
TREAT HF in DMF (2.00 M) were 18 mL h−1 and 1.5 mL h−1,
b
respectively. 19F NMR yield determined with an internal standard.
derived CuCF3 that are presented in Table 1 and discussed
throughout this report.
Initially, the flow rates of 1 in DMF and fluoroform were set
to 18 mL h−1 and 6 mL min−1, respectively, in order to
maintain the CHF3 to Cu ratio at 2:1 (Table 1, entries 1−4).
The residence time (RT), i.e. the reaction time before the
addition of TREAT HF, was controlled by the length of the
reaction zone, a 1.55 mm inner diameter PTFE (polytetra-
fluoroethylene; Teflon) tube. As pictorially shown in Figure 1,
the volume of the fluoroform gas decreased as it moved inside
the cupration reaction zone due to dissolution and reaction
with 1. Changing RT from 1.25 to 2.00 and then to 2.50 min
led to an increase in the yield of CuCF3 from 85% to 87%, and
to 94%, respectively (entries 1−3); a further increase in RT to
2.75 min, however, did not affect the yield (entry 4). Although
the best yield of 94% obtained in this series of experiments
(entries 3 and 4) was as high as that from the batch process,8
further optimization was performed in order to minimize the
amount of CHF3 used. Toward this goal, the fluoroform feed
rate was lowered to 4.5 mL min−1 to change the CHF3 to Cu
B
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 2. Four different configurations for the trifluoromethylation reactions with the CuCF3 reagent from the continuous flow process.
ratio to 1.5. By subsequently varying RT in a 6.00−11.33 min
range (entries 5−8) we found the optimal value of 9.33 min for
the formation of CuCF3 in the highest possible yield of 94%
(entry 7). Still high yet slightly lower yields (88−92%) resulted
from both shorter (entries 5 and 6) and longer (entry 8)
reaction times. Finally, a series of runs was performed with a
CHF3 to Cu ratio of 1.2 that was achieved by slowing the
fluoroform feed rate to 3.6 mL min−1. In these experiments, the
yield of CuCF3 passed through a maximum (86%; entry 10) at
RT = 8.83 min, an intermediate value between the other two
points (entries 9 and 11). The lower CHF3 concentration in
these three runs resulted in slower cupration rates, and hence
the CuCF3 product began to decompose (see above)8 at a
noticeable extent even before full conversion could be reached.
While at the shorter RT of 5.58 min (entry 9) the yield of
CuCF3 was only 75% because of incomplete conversion, the
longer RT of 11.83 min (entry 11) resulted in partial
decomposition of the product, resulting in 79% yield. It is
strongly believed, however, that an engineering solution is
available for the cupration of CHF3 in a continuous manner in
over 90% yield while using CHF3 in only 20% excess and even
less. One option is to run the process at a slightly increased
pressure in the reaction zone. Another possibility is to recycle
and reuse the easily separable fluoroform gas from the head
space of the receiver during the process employing CHF3 in the
1.5-fold excess that is sufficient to obtain CuCF3 in 94% yield
under normal pressure (entry 7). Conducting such studies,
however, was beyond the scope of our work.
In all of the experiments described above, the stabilization of
CuCF3 was performed by adding TREAT HF to the receiving
vessel, as shown in Figure 1. An alternative setup with a T-
connector at the end of the cupration reaction zone for TREAT
HF addition was also tested but found less efficient (∼70%
yield). The inferior results were caused by the KF precipitate
produced in the stabilization step, thereby preventing the free
passage of the reaction mixture through the small-diameter tube
and eventually clogging the system. Although increasing the
reactor tube diameter could eliminate this problem, such
further developments are the subject of a separate project.
Trifluoromethylation Reactions with CuCF3 Produced
in the Flow Process. Having developed and optimized the
continuous flow synthesis of CuCF3 from fluoroform, we
examined the usefulness of the thus prepared reagent for a
series of known trifluoromethylation reactions. In these studies
we used the temperature and reactant ratios previously found
and optimized for the analogous transformations9−11 per-
formed with the CuCF3 reagent from the batch process.8 The
different setups used in the current studies are shown in Figure
2, and a summary of the results is presented in Table 2. All
yields were determined by 19F NMR since isolation procedures
for the products forming in these reactions have been
developed and reported previously.9−11
First, the trifluoromethylation reaction of 2-bromoacetophe-
none (2a)11 was examined using configuration (i). In that
experiment (Table 2, entry 1), TREAT HF was employed in
superstoichiometric quantities (0.63 mol per mol Cu; see the
Experimental Section for specifics). Of the total of 0.63 mol of
TREAT HF used, 0.33 mol (1 equiv) was required for the
conventional stabilization of the copper reagent (Scheme 1),
whereas the extra 0.30 mol was needed to avoid the HF
elimination from the desired trifluoromethylated product.11 As
the CuCF3 was produced and stabilized with TREAT HF in
excess, a solution of 2a was fed simultaneously to the glass
receiver/reactor at such a rate so as to maintain the CuCF3 to
substrate ratio at 1.2. Under such conditions, 2a was
trifluoromethylated in 72−73% yield (Table 2, entry 1). This
yield was noticeably lower than that (93−96%) previously
reported11 for the same reaction with the CuCF3 reagent from
the batch process. In the latter, however, the CuCF3 was first
prepared and stabilized with only 1 equiv of HF from TREAT
HF and then used for the reaction in the presence of an
additional amount of TREAT HF.11 The lower yields obtained
under the current flow conditions were caused by the enhanced
side formation of the hydrodebromination product, acetophe-
none (∼12% by GC−MS). This undesired hydrodebromina-
tion was likely prompted by the reasonably DMF-soluble
KHF2, generated from extra HF and KF formed upon addition
of superstoichiometric amounts of TREAT HF to the just
produced unstabilized CuCF3 solution. In contrast, the
stabilization with stoichiometric quantities of TREAT HF
produces only KF that is poorly soluble in DMF and efficiently
precipitates out on standing. In this way, virtually potassium-
free CuCF3 solutions are produced,8 to which extra TREAT HF
can be added11 without the risk of KHF2 formation. The
C
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 2. Trifluoromethylation of different substrates with fluoroform-derived CuCF3 produced in the flow process
a
b
All reactions were performed at 23 °C, using CuCF3 prepared in the optimized flow process (Table 1, entry 7). Reaction time after mixing the
c
d
e
f
substrate with CuCF3. Yields determined by 19F NMR. See Figure 2 for setup configurations. No extra TREAT HF used. Reactions performed in
air.
trifluoromethylation of α-haloketones with such fluoroform-
derived CuCF3 solutions containing extra TREAT HF, but not
KHF2, is significantly higher yielding and chemoselective,
producing only very small quantities, if any, of the hydro-
dehalogenation side-products.11 However, a substantially more
complex flow setup would be required to perform the
trifluoromethylation of α-haloketones with such KHF2-free,
TREAT HF-enriched CuCF3 reagent in a fully continuous
manner.
of 2a performed with the two-step addition of TREAT HF in
configurations (ii)−(iv) did not exceed 4−6%, whereas the all-
at-once TREAT HF addition method gave ∼12% of the
hydrodebrominated side-product.
The results described above indicated that configurations (ii)
and/or (iii) should be used for further runs, in which we
intended to trifluoromethylate aromatic substrates. Thus, 2-
iodobenzamide (2b) and 2-bromobenzoic acid (2c) were
trifluoromethylated using configuration (ii) to give 3b and 3c in
79−85% and 80−89% yield, respectively (Table 2, entries 5 and
6). These yields are nearly identical with those (82−87%)
previously reported for the same transformations employing the
CuCF3 reagent from the batch preparation.9 It is noteworthy
that no extra TREAT HF was needed for the reaction of 2c that
bears an acidic functionality (COOH). Finally, we performed
oxidative trifluoromethylation of 2-naphthylboronic acid (2d)
and 4-biphenylylboronic acid (2e) with CuCF3 produced in the
continuous process. As in the reported10 batch method, both
reactions were performed in open air, without extra TREAT
HF to give 3d and 3e quantitatively (entries 7 and 8).
Considering the above, we tested the three other
configurations (ii), (iii), and (iv) shown in Figure 2 for the
trifluoromethylation of 2a. In all these experiments, the CuCF3
stabilized with 0.33 mol of TREAT HF per 1 mol of Cu was
prepared using the flow reactor (Figure 1), as described above.
The resultant reagent solution was then transferred with a
peristaltic pump to a glass reactor for the trifluoromethylation
to be performed. Using configuration (ii), the required extra
amount of TREAT HF, 2a, and the already stabilized CuCF3
were fed into the reactor simultaneously. In this way, 3a was
produced in 85−87% yield (Table 2, entry 2), nearly as high as
the best yields reported previously for the batch method.11
With configuration (iii), extra TREAT HF (0.2 equiv) was first
added to the already stabilized CuCF3, which was then used for
the reaction with 2a. This technique afforded 3a in a yield (81−
83%) that was indistinguishable, within the determination error,
from the one achieved in the setup (ii) experiment (entries 2
and 3). It is worth noting that the error of yield determination
by 19F NMR is at least 5%. Performing the trifluoromethy-
lation of 2a in a continuous manner with setup (iv) using a T-
connection (Figure 2) resulted in a lower yield of 3a (75%,
entry 4) because of precipitation of inorganic copper
byproducts and eventual clogging of the tube. As expected,
the side formation of acetophenone in the trifluoromethylations
CONCLUSIONS
■
We have successfully demonstrated the first continuous flow
process for the synthesis of the valuable trifluoromethylating
reagent CuCF3 via the direct cupration of fluoroform. The
reaction smoothly occurs at room temperature and atmospheric
pressure. Under optimized conditions, the yield of CuCF3 in
the flow reactor (up to 94%) is as high as in the previously
developed batch process. The continuously produced CuCF3
has been found to trifluoromethylate a number of organic
substrates with excellent efficiencies and yields comparable to
those previously observed for the same reagent prepared by the
D
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
batch methodology. It is hoped that the demonstrated
feasibility of the flow synthesis of the low-cost fluoroform-
derived CuCF3 adds to its attractiveness as a trifluoromethylat-
ing reagent for larger-scale operations.16
evacuated to ∼1−2 mbar and backfilled with argon three times,
the flow synthesis was commenced. As described above, the
yield of CuCF3 in the first 2−3 min of the run was always
lower, and the product solution was dark because of partial
decomposition with trace residual air in the system. After this
initial period, the solution produced was pale yellow, and the
CuCF3 yield was consistently high (see Table 1 for specifics), as
established by frequent 19F NMR monitoring (every 5−10
min).
EXPERIMENTAL SECTION
■
General Information. Fluoroform (98%) in a cylinder was
purchased from Apollo Scientific. All other chemicals and
solvents were purchased from Aldrich, Alfa Aesar, TCI, and
Acros chemical companies. Anhydrous DMF (Alfa Aesar or
from an MBraun SPS) was stored over freshly calcined 4 Å
molecular sieves in a glovebox. All manipulations were
performed under argon, unless noted otherwise. NMR spectra
were obtained with a Bruker Avance 400 Ultrashield NMR
spectrometer. Quantitative 19F NMR analyses were carried out
with D1 = 5 s. An Agilent Technologies 7890A chromatograph
equipped with a 5975C MSD unit was used for GC−MS
analysis.
3,3,3-Trifluoro-1-phenylpropan-1-one (3a; Table 2,
entries 1−4). Procedure A (entry 1). Configuration (i)
shown in Figure 2 was used. All reaction solutions for the
synthesis were prepared inside a glovebox and placed in
syringes equipped with PTFE adapters fitted with 1/8 in. ball
valves. A 5 mL syringe was charged with 2.9 mL of a DMF
solution of 2-bromoacetophenone (2a; 995 mg; 5.00 mmol;
[2a] = 1.72 M) and 1,3-bis(trifluoromethyl)benzene (194 μL;
1.25 mmol; internal standard). A 20 mL syringe was filled up
with a 0.48 M solution of 1 in DMF prepared as described
above. A 10 mL syringe was charged with 5 mL of a 2 M
TREAT HF solution prepared as described above. The three
syringes were brought out and connected to the reaction
system of configuration (i) shown in Figure 2, and the
continuous cupration process was commenced using the
parameters specified in entry 7 of Table 1, except that the
rate of addition of the TREAT HF solution was 2.9 mL h−1.
The first ∼2 mL portion of the dark, partially oxidized CuCF3
solution product was completely removed from the receiver/
reactor and discarded. 19F NMR analysis of the fully withdrawn
next, light-yellow 1 mL fraction of the product indicated 90%
yield of CuCF3. At that point, the addition of the solution of 2a
was started at a rate of 3.87 mL h−1 as the CuCF3 was
continuously produced. After ∼3 mL of the reaction mixture
accumulated in the reactor, the entire amount was withdrawn
via syringe, kept under argon for an additional 15 min, and
treated with ∼10 mL of water and ∼4 mL of ether in air. After
agitation, the organic layer was separated and filtered through a
short silica gel plug. Quantitative 19F NMR analysis of the
filtrate indicated that 3a was produced in 73% yield. 19F NMR,
δ: = −61.2 (t, J = 10.3 Hz).11 Acetophenone (∼11%) was
detected in the filtrate by GC−MS. The continuously produced
reaction mixture was withdrawn from the reactor two more
times. Analysis of these two fractions as described above
indicated 73% and 72% yield of 3a (19F NMR) and ∼12% and
∼12% of acetophenone (GC−MS).
Procedure B (entry 2). Configuration (ii) shown in Figure 2
was used. CuCF3 in DMF was prepared as described in Table 1
(entry 7) in 92% yield. After the KF byproduct had settled at
the bottom of the receiver, the solid-free supernatant was
pumped into the glass reactor with the peristaltic pump through
a PTFE tube equipped with a microfiber filter at the inlet.
Simultaneously, solutions of 2a and TREAT HF in DMF were
fed to the same reactor via two independently operating syringe
pumps. A 5 mL syringe was used for the addition of the
substrate solution (2.8 mL) in DMF containing 2a (995 mg;
5.00 mmol; [2a] = 1.79 M) and 1,3-bis(trifluoromethyl)-
benzene (194 μL; 1.25 mmol; internal standard). The TREAT
HF solution (2 M; 4 mL) was added via a 10 mL syringe. The
rates of addition of the CuCF3, 2a, and TREAT HF solutions
were 19 mL h−1, 3.34 mL h−1, and 0.9 mL h−1, respectively.
After the first ∼3 mL portion was discarded, three sequential
fractions of ∼3 mL each were withdrawn and analyzed by 19F
Continuous Reaction System. Gaseous fluoroform was
fed into the system (Figure 1) by a thermal mass flow
controller (MFC, Alicat Scientific) calibrated for CHF3. A
PTFE tube of 3.25 mm outer diameter and 1.55 mm inner
diameter was used as both the reactor and connector between
the syringes and valves with standard SS-316 compression
fittings (Hy-Lok). The MFC outlet was connected to an SS-316
T-piece where the solution of 1 and gaseous CHF3 met and
were mixed together. An in-house fabricated PTFE adapter
facilitated the coupling of the syringe to the reaction system. To
keep solutions of 1 and TREAT HF in syringes isolated from
oxygen and moisture during the transfer from the glovebox to
the reaction system, a standard 1/8 in. ball valve was fitted on
the PTFE adapter. The liquid reagents were transferred to a
syringe of appropriate volume and fed to the reaction system
using high-accuracy syringe pumps (Fisher Scientific). The
outlet streams were connected via PTFE-to-glass fittings (Bola,
Germany) to a glass receiver/reactor for the stabilization with
TREAT HF (Step 3). The receiver/reactor was equipped with a
stopcock for evacuation/backfilling with argon, an airlock to
maintain a slow argon flow throughout the experiment, and a
PTFE-coated magnetic stir bar for continuous agitation. A
TREAT HF solution was added to the receiver/reactor using a
separate syringe pump. In the trifluoromethylation experiments,
the CuCF3 solution product was pumped out continuously
from the glass receiver/reactor using a peristaltic pump and fed
to the second glass reactor or premixed with the substrate,
depending on the configuration employed (see above, Figure 2
and Table 2). Vigorous agitation of the trifluoromethylation
reaction mixtures was used in all experiments.
Preparation of [K(DMF)][Cu(OBu-t)2] (1)8 and the
Continuous Cupration of Fluoroform. In a glovebox, a
mixture of CuCl (purity 99%; 1.50 g; 15.1 mmol), t-BuOK
(purity 97%; 3.54 g; 30.6 mmol), and DMF (25 mL) was
vigorously stirred for 30 min at room temperature. The
precipitated KCl was filtered off and washed on the filter with
DMF (5 mL). To the combined filtrate and the washing, was
added PhF (purity 99%; 0.42 mL; 4.5 mmol) as an internal
standard for CuCF3 yield determination. The thus prepared
solution was placed in a 20 mL syringe. A 2 M solution of
TREAT HF in DMF prepared inside the glovebox from 3.3 mL
(20 mmol) of TREAT HF in a 10 mL graduated flask was
placed in a 10 mL syringe. Both syringes equipped with PTFE
adapters fitted with 1/8 in. ball valves were brought out and
connected to the continuous flow reactor. After the system was
E
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
NMR and GC−MS as described above. The yields of
3a(acetophenone) were 85(5)%, 87(5)%, and 85(4)%.
yields of 3b were 85%, 83%, and 79%. 19F NMR, δ: −58.3 (s).9
After treatment of the samples with H2O/ether, the organic
layers were analyzed by GC−MS. No side products were
detected. A small quantity of the starting material (2b) was
detected in the third fraction but not in the first two.
Procedure C (entry 3). CuCF3 in DMF was prepared as
described in Table 1 (entry 7) in 91% yield. After the KF
byproduct had settled at the bottom of the receiver, the solid-
free supernatant was pumped into the glass reactor with the
peristaltic pump through a PTFE tube equipped with a
microfiber filter at the inlet. Simultaneously, solutions of 2a and
TREAT HF in DMF were fed to the system via two
independently operating syringe pumps, as shown in Figure
2, setup (iii). A 5 mL syringe was used for the addition to the
reactor of the substrate solution (2.8 mL) in DMF containing
2a (995 mg; 5.00 mmol; [2a] = 1.79 M) and 1,3-
bis(trifluoromethyl)benzene (194 μL; 1.25 mmol; internal
standard). The TREAT HF solution (2 M; 4 mL) was added
via a 10 mL syringe into the CuCF3 solution flow. The rates of
addition of the CuCF3, 2a, and TREAT HF solutions were 19
mL h−1, 3.30 mL h−1, and 0.9 mL h−1, respectively. After the
first ∼3 mL portion was discarded, three sequential fractions of
∼3 mL each were withdrawn and analyzed by 19F NMR and
GC−MS as described above. The yields of 3a(acetophenone)
were 81(4)%, 80(5)%, and 83(5)%.
2-(Trifluoromethyl)benzoic acid (3c; Table 2, entry 6).
Configuration (ii) or (iii) shown in Figure 2 was used (no extra
TREAT HF). CuCF3 in DMF was prepared as described in
Table 1 (entry 7) in 88% yield. After the KF byproduct had
settled at the bottom of the receiver, the solid-free supernatant
was pumped into the glass reactor with the peristaltic pump
through a PTFE tube equipped with a microfiber filter at the
inlet. Simultaneously, a solution of 2c was fed to the same
reactor via a syringe pump. A 5 mL syringe was used for the
addition of the substrate solution (2.4 mL) in DMF containing
2c (503 mg; 2.50 mmol; [2c] = 1.04 M) and 1,3-
bis(trifluoromethyl)benzene (194 μL; 1.25 mmol; internal
standard). The rates of addition of the CuCF3 and 2c solutions
were 19 mL h−1, and 6.49 mL h−1, respectively. Four sequential
fractions of ∼3 mL each were withdrawn and analyzed by 19F
NMR “as is”. The yields of 3c were 80%, 88%, 88%, and 89%.
19F NMR, δ: −58.4 (s).9
Procedure D (entry 4). CuCF3 in DMF was prepared as
described in Table 1 (entry 7) in 94% yield. After the KF
byproduct had settled at the bottom of the receiver, the solid-
free supernatant was pumped into the glass reactor with the
peristaltic pump through a PTFE tube equipped with a
microfiber filter at the inlet. Simultaneously, solutions of 2a and
TREAT HF in DMF were fed to the flow of the CuCF3 reagent
solution via two independently operating syringe pumps, as
shown in Figure 2, setup (iv). A 5 mL syringe was used for the
addition of the substrate solution (2.8 mL) in DMF containing
2a (995 mg; 5.00 mmol; [2a] = 1.79 M) and 1,3-
bis(trifluoromethyl)benzene (194 μL; 1.25 mmol; internal
standard). The TREAT HF solution (2 M; 4 mL) was added
via a 10 mL syringe into the CuCF3 solution flow. The rates of
addition of the CuCF3, 2a, and TREAT HF solutions were 19
mL h−1, 3.49 mL h−1, and 0.9 mL h−1, respectively. As the
reaction occurred in the tube, a solid was produced. Three
sequential fractions of ∼3 mL each were withdrawn from the
reactor and analyzed by 19F NMR and GC−MS as described
above. The yields of 3a(acetophenone) were 71(6)%, 75(5)%,
and 74(4)%. Also, unreacted 2a (∼5%) was detected in the first
fraction by GC−MS. A few min after the third fraction was
withdrawn, signs of clogging of the system with the solid
byproduct were observed and the experiment was stopped.
2-(Trifluoromethyl)benzamide (3b; Table 2, entry 5).
Configuration (ii) shown in Figure 2 was used. CuCF3 in DMF
was prepared as described in Table 1 (entry 7) in 88% yield.
After the KF byproduct had settled at the bottom of the
receiver, the solid-free supernatant was pumped into the glass
reactor with the peristaltic pump through a PTFE tube
equipped with a microfiber filter at the inlet. Simultaneously,
solutions of 2b and TREAT HF in DMF were fed to the same
reactor via two independently operating syringe pumps. A 5 mL
syringe was used for the addition of the substrate solution (2.8
mL) in DMF containing 2b (617 mg; 2.50 mmol; [2b] = 0.89
M) and 1,3-bis(trifluoromethyl)benzene (194 μL; 1.25 mmol;
internal standard). The TREAT HF solution (2 M; 4 mL) was
added via a 10 mL syringe. The rates of addition of the CuCF3,
2b, and TREAT HF solutions were 19 mL h−1, 7.77 mL h−1,
and 0.9 mL h−1, respectively. Three sequential fractions of ∼3
mL each were withdrawn and analyzed by 19F NMR “as is”. The
2-(Trifluoromethyl)naphthalene (3d; Table 2, entry 7).
Configuration (ii) or (iii) shown in Figure 2 was used (no extra
TREAT HF). CuCF3 in DMF was prepared as described in
Table 1 (entry 7) in 89% yield. After the KF byproduct had
settled at the bottom of the receiver, the solid-free supernatant
was pumped into the glass reactor with the peristaltic pump
through a PTFE tube equipped with a microfiber filter at the
inlet. Simultaneously, a solution of 2d (430 mg; 2.50 mmol)
and 4,4′-difluorobiphenyl (178 mg; 0.94 mmol; internal
standard) in a mixture of DMF (1.2 mL) and toluene (2.0
mL) of 3.8 mL total volume ([2d] = 0.66 M) was fed to the
reactor in air via a syringe pump. The rates of addition of the
CuCF3 and 2d solutions were 19 mL h−1 and 5.68 mL h−1,
respectively. Three sequential fractions of ∼3 mL each were
withdrawn. Each fraction was vigorously agitated in air for 1 h
and then treated with H2O/ether, and the organic layers were
analyzed by 19F NMR. The yields of 3d were 92%, 99%, and
99%. 19F NMR, δ: −62.7 (s).10 The yield of naphthalene in
these fractions was 4%, 2%, and 1%, respectively (GC−MS).
4-(Trifluoromethyl)biphenyl (3e; Table 2, entry 8).
Configuration (ii) or (iii) shown in Figure 2 was used (no extra
TREAT HF). CuCF3 in DMF was prepared as described in
Table 1 (entry 7) in 91% yield. After the KF byproduct had
settled at the bottom of the receiver, the solid-free supernatant
was pumped into the glass reactor with the peristaltic pump
through a PTFE tube equipped with a microfiber filter at the
inlet. Simultaneously, a solution of 2e (495 mg; 2.50 mmol)
and 4,4′-difluorobiphenyl (178 mg; 0.94 mmol; internal
standard) in a mixture of DMF (1.0 mL) and toluene (3.4
mL) of 4.9 mL total volume ([2e] = 0.51 M) was fed to the
reactor in air via a syringe pump. The rates of addition of the
CuCF3 and 2e solutions were 19 mL h−1 and 7.54 mL h−1,
respectively. Four sequential fractions of ∼3 mL each were
withdrawn. Each fraction was vigorously agitated in air for 0.5 h
and then treated with H2O/ether, and the organic layers were
analyzed by 19F NMR. The yields of 3e were 99%, 98%, 99%,
and 98%. 19F NMR, δ: −62.8 (s).10 The yield of biphenyl could
not be estimated in this experiment because of the close
retention times of Ph2 and 4,4′-difluorobiphenyl (internal
standard) under the GC−MS conditions used. Therefore, the
F
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(8) Zanardi, A.; Novikov, M. A.; Martin, E.; Benet-Buchholz, J.;
Grushin, V. V. J. Am. Chem. Soc. 2011, 133, 20901.
trifluoromethylation was repeated in the absence of 4,4′-
difluorobiphenyl. The yield of biphenyl was estimated at ≤1%.
(9) Lishchynskyi, A.; Novikov, M. A.; Martin, E.; Escudero-Adan
C.; Novak, P.; Grushin, V. V. J. Org. Chem. 2013, 78, 11126.
(10) Novak, P.; Lishchynskyi, A.; Grushin, V. V. Angew. Chem., Int.
Ed. 2012, 51, 7767.
(11) Novak, P.; Lishchynskyi, A.; Grushin, V. V. J. Am. Chem. Soc.
́
, E.
́
AUTHOR INFORMATION
■
́
Corresponding Authors
*E-mail for A.U.: aurakawa@iciq.es.
*E-mail for V.V.G.: vgrushin@iciq.es.
́
2012, 134, 16167.
(12) (a) See, for example: Chemical Reactions and Processes under
Flow Conditions; Luis, S. V., Garcia-Verdugo, E., Eds.; RSC:
Cambridge, U.K., 2010. (b) For a recent, excellent review of flow
microreactor synthesis in organofluorine chemistry, see: Amii, H.;
Nagaki, A.; Yoshida, J.-i. Beilstein J. Org. Chem. 2013, 9, 2793.
(13) Konovalov, A. I.; Benet-Buchholz, J.; Martin, E.; Grushin, V. V.
Angew. Chem., Int. Ed. 2013, 52, 11637.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank The ICIQ Foundation and the Spanish Government
(Grant CTQ2011-25418) for support of this work. A.B. thanks
AGAUR-Generalitat de Catalunya, Spain, for the FI grant.
(14) Lishchynskyi, A.; Grushin, V. V. J. Am. Chem. Soc. 2013, 135,
12584.
(15) Although inclusion of Step 1 in the overall flow setup was
beyond the scope of the current work, our preliminary experiments
indicated that this is possible.
REFERENCES
■
(1) For selected monographs, see: (a) Hudlicky, M. Chemistry of
Organic Fluorine Compounds; Ellis Horwood: New York, 1976.
(b) Banks, R. E. Organofluorine Chemicals and Their Industrial
Applications; Ellis Horwood: West Sussex, U.K., 1979. (c) Filler, R.;
Kobayashi, Y. Biomedicinal Aspects of Fluorine Chemistry; Kodansha-
Elsevier: New York, 1982. (d) Clark, J. H.; Wails, D.; Bastock, T. W.
Aromatic Fluorination; CRC Press: Boca Raton, FL, 1996. (e) Kirsch,
P. Modern Fluoroorganic Chemistry; Wiley-VCH: Weinheim, 2004.
(f) Uneyama, K. Organofluorine Chemistry; Blackwell: Oxford, U. K.,
2006. (g) Ojima, I. Fluorine in Medicinal Chemistry and Chemical
Biology; Wiley-Blackwell: Chichester, U. K., 2009. (h) Petrov, V. A.
Fluorinated Heterocyclic Compounds. Synthesis, Chemistry and Applica-
tions; Wiley: Hoboken, NJ, 2009.
(2) For selected reviews, see: (a) Burton, D. J.; Yang, Z. Y.
Tetrahedron 1992, 48, 189. (b) McClinton, M. A.; McClinton, D. A.
Tetrahedron 1992, 48, 6555. (c) Umemoto, T. Chem. Rev. 1996, 96,
1757. (d) Burton, D. J.; Lu, L. Top. Curr. Chem. 1997, 193, 45.
(e) Prakash, G. K. S.; Yudin, A. K. Chem. Rev. 1997, 97, 757.
(f) Guittard, F.; Taffin de Givenchy, E.; Geribaldi, S.; Cambon, A. J.
Fluorine Chem. 1999, 100, 85. (g) Singh, R. P.; Shreeve, J. M.
Tetrahedron 2000, 56, 7613. (h) Jeschke, P. ChemBioChem. 2004, 5,
570. (i) Ma, J.-A.; Cahard, D. Chem. Rev. 2004, 104, 6119.
(j) Schlosser, M. Angew. Chem., Int. Ed. 2006, 45, 5432. (k) Ma, J.-
A.; Cahard, D. J. Fluorine Chem. 2007, 128, 975. (l) Uneyama, K.;
Katagiri, T.; Amii, H. Acc. Chem. Res. 2008, 41, 817. (m) Ma, J.-A.;
Cahard, D. Chem. Rev. 2008, 108, PR1. (n) Shibata, N.; Matsnev, A.;
Cahard, D. Beilstein J. Org. Chem. 2010, DOI: 10.3762/bjoc.6.65.
(o) Sato, K.; Tarui, A.; Omote, M.; Ando, A.; Kumadaki, I. Synthesis
2010, 1865. (p) Dhara, M. G.; Banerjee, S. Prog. Polym. Sci. 2010, 35,
1022. (q) Roy, S.; Gregg, B. T.; Gribble, G. W.; Le, V.-D.; Roy, S.
Tetrahedron 2011, 67, 2161. (r) Tomashenko, O. A.; Grushin, V. V.
Chem. Rev. 2011, 111, 4475. (s) Qing, F.-L.; Zheng, F. Synlett 2011,
1052. (t) Dilman, A. D.; Levin, V. V. Eur. J. Org. Chem. 2011, 831.
(u) Nie, J.; Guo, H.-C.; Cahard, D.; Ma, J.-A. Chem. Rev. 2011, 111,
455. (v) Grygorenko, O. O.; Artamonov, O. S.; Komarov, I. V.;
(16) A comment is due on some recently published incorrect
assertions regarding our CHF3 cupration reaction and its product. In
one paper, it has been stated that our new fluoroform-based synthetic
methods8−11 “require excess of CF3H [2 to 3 equivalent (equiv)] and
have limited scope”.17 Aside from the scope of our methods4,8−11,13,14
that is a matter of personal judgment, we are unaware of the source of
the wrong information that 2−3 equiv of CHF3 are required to obtain
the CuCF3 reagent. Our data previously obtained for the batch
cupration process8 and the current results presented in Table 1
(entries 5−11) clearly indicate that the CuCF3 can be obtained in up
to 86−94% yield using only 1.2−1.5 equiv of CHF3. In another recent
report,18 it is stated “Although the functional group tolerance and
yields of this [our] method are high, the CuCF3 reagent cannot be
stored. Thus, each reaction must be initiated by generation of CuCF3
from gaseous HCF3, and such a transformation is challenging to
conduct in common laboratory settings.” This statement is particularly
surprising in light of the fact that in our 2011 report8 that is cited in ref
18, it is clearly written that “On storage of the solution at room
temperature, a minor drop in the original yield of 95% to 93% and
91% was detected after 24 h and 3 days, respectively. An aliquot of the
same solution that was stored at −35 °C for 8 days showed no sign of
decomposition.”8 Furthermore, as mentioned above, stabilized
solutions of fluoroform-derived CuCF3 can be stored at −30 °C
without decomposition for months.
(17) Prakash, G. K. S.; Jog, P. V.; Batamack, P. T. D.; Olah, G. A.
Science 2012, 338, 1324.
(18) Mormino, M. G.; Fier, P. S.; Hartwig, J. F. Org. Lett. 2014, 16,
1744.
Mykhailiuk, P. K. Tetrahedron 2011, 67, 803. (w) Acena, J. L.;
̃
Sorochinsky, A. E.; Soloshonok, V. A. Synthesis 2012, 44, 1591.
(x) Mace,
́
Y.; Magnier, E. Eur. J. Org. Chem. 2012, 2479. (y) Chen, P.;
, P.;
Liu, G. Synthesis 2013, 45, 2919. (z) Lishchynskyi, A.; Novak
́
Grushin, V. V. In Science of Synthesis: C-1 Building Blocks in Organic
Synthesis 2; van Leeuwen, P. W. N. M., Ed.; Thieme: Stuttgart, 2013;
pp 367−408.
(3) For a very rare example of trifluoromethylation on a kilogram
scale, see: Mulder, J. A.; Frutos, R. P.; Patel, N. D.; Qu, B.; Sun, X.;
Tampone, T. G.; Gao, J.; Sarvestani, M.; Eriksson, M. C.; Haddad, N.;
Shen, S.; Song, J. J.; Senanayake, C. H. Org. Process Res. Dev. 2013, 17,
940.
(4) Grushin, V. V. Chim. Oggi - Chem. Today 2014, 32, 81.
(5) McCulloch, A.; Lindley, A. A. Atmos. Environ. 2007, 41, 1560.
(6) Han, W.; Li, Y.; Tang, H.; Liu, H. J. Fluorine Chem. 2012, 140, 7.
(7) Bomgardner, M. M. Chem. Eng. News 2013, 91 (26), 6.
G
dx.doi.org/10.1021/op500109v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX