Synthesis of Celecoxib, Mavacoxib, SC-560, Fluxapyroxad, and Bixafen Enabled by Continuous Flow Reaction Modules

Article abstract of DOI:10.1002/ejoc.201700992

Multi-step continuous flow synthesis enables a parallel approach to obtain agrochemicals and pharmaceuticals containing 3-fluoroalkyl pyrazole cores. In this system, fluorinated amines are transformed into pyrazole cores through a telescoped in situ generation and consumption of diazoalkanes. Once synthesized, additional continuous flow and batch reactions add complexity to the pyrazole core via C–N arylation and methylation, TMS cleavage, and amidation. Using this modular assembly line approach, Bixafen and Fluxapyroxad were synthesized in 38 % yield over four continuous flow steps in an overall reaction time of 56 min. Finally, coupling selected continuous flow processes with an offline (batch) Ullmann coupling afforded Celecoxib, Mavacoxib, and SC-560 in 33–54 % yield over two to three steps.

Full text of DOI:10.1002/ejoc.201700992

DOI: 10.1002/ejoc.201700992
Full Paper
Continuous Flow Synthesis
Synthesis of Celecoxib, Mavacoxib, SC-560, Fluxapyroxad, and
Bixafen Enabled by Continuous Flow Reaction Modules
Joshua Britton[a] and Timothy F. Jamison*[a]
Abstract: Multi-step continuous flow synthesis enables a paral- methylation, TMS cleavage, and amidation. Using this modular
lel approach to obtain agrochemicals and pharmaceuticals con- assembly line approach, Bixafen and Fluxapyroxad were synthe-
taining 3-fluoroalkyl pyrazole cores. In this system, fluorinated sized in 38 % yield over four continuous flow steps in an overall
amines are transformed into pyrazole cores through a tele- reaction time of 56 min. Finally, coupling selected continuous
scoped in situ generation and consumption of diazoalkanes. flow processes with an offline (batch) Ullmann coupling af-
Once synthesized, additional continuous flow and batch reac- forded Celecoxib, Mavacoxib, and SC-560 in 33–54 % yield over
tions add complexity to the pyrazole core via C–N arylation and two to three steps.
Introduction
addition, C–N arylation, C–N methylation, TMS removal and
amidation.[6] Herein, these continuous flow modules are further
optimized and subsequently utilized to rapidly synthesize
popular agrochemicals Fluxapyroxad and Bixafen,[7] and cou-
Continuous flow has emerged as an enabling technology in
organic synthesis over the last decade.[1] The growth of continu-
ous flow can be credited to its benefits over batch chemistry. In
continuous flow systems, solvents and reactants can be heated
above their atmospheric boiling points to afford improved ki-
netics,[2] hazardous reactions can be confined to small volume
reactor coils for improved safety,[1e,3] and reaction scale can be
addressed through parallel processing to meet a demand.[4]
In multi-step continuous flow synthesis, sequential reactor
coils are used to accrue molecular complexity.[1a,1f] While a
large number of valuable multi-step continuous flow syntheses
exist, only a few can be considered modular, or “assembly line
processes”.[5] The concept of assembly lines originates from
Henry Ford's achievements in the motor industry. Here, simple
frameworks are passed along a conveyer belt where additional
pieces are installed in a specific order to create a car. In a con-
tinuous flow sense, passing molecules through specific reactor
modules dictates the chemical functionality present in the final
product. Additionally, modules often have broad substrate
compatibility, and reconfiguring the sequence of modules can
enable the synthesis of different target compounds. After re-
cently achieving an assembly line synthesis of Measles thera-
peutic AS-136A through this approach, we hypothesized that
several other bioactive molecules could benefit.[6]
pling continuous flow and batch techniques allowed
a
parallel synthesis of non-steroidal anti-inflammatory drugs
(NSAIDs) Celecoxib (Celebrex™),[8] Mavacoxib (Trocoxil™),[9] and
SC-560[8,10] (Figure 1). Furthermore, the discovery of a radical
pathway for the synthesis of a symmetrical vinyl ether, and a
lipase-catalyzed hydrolysis of a pyrazole ester is disclosed.
In our initial communication, six continuous flow modules
were reported that mediate diazoalkane formation, [3+2] cyclo-
[a] Department of Chemistry, Massachusetts Institute of Technology,
77 Massachusetts Ave., Cambridge, MA 02139, USA
E-mail: tfj@mit.edu
Figure 1. Pharmaceuticals and agrochemicals containing CF3 and CF2H groups
at the 3-position. Molecules are colored to show structural similarities. In our
initial communication, AS-136A was synthesized by a continuous flow assem-
bly line strategy. Here, Fluxapyroxad, Bixafen, Mavacoxib, Celecoxib, and
SC-560 are synthesized by an advanced assembly line strategy.
web.mit.edu/chemistry/jamison
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under https://doi.org/10.1002/ejoc.201700992.
Eur. J. Org. Chem. 0000, 0–0
1
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
methyl diazomethane 1, or 2,2,2-trifluoromethyl diazomethane
2 with an alkyne (Figure 3). In optimizing this transformation, 1
was synthesized in situ for immediate consumption in a
[3+2] cycloaddition with ethyl propiolate. The synthesis of 1
(Module 1) requires only 5 min at 60 °C (Figure 3B and D), and
[2+3] cycloaddition (Module 2) requires only 10 min at 90 °C
(Figure 3C and E). Interestingly, any deviation away from these
optimal reaction conditions significantly decreased the yield.
For example, increasing the temperature of module 1 passed
60 °C results in lower yields due to diazoalkane and tBuONO
decomposition (Figure 3B),[11] while operating module 2 passed
90 °C lead to rapid gas evolution with concomitant lower yields
(Figure 3C). With these optimized conditions, 3 and 4 were af-
forded in 92 % and 99 % yields respectively.
In an assembly line synthesis, each reactor module performs a
dedicated transformation that is part of an overall sequence. To
enable the parallel synthesis of a range of pyrazole products,
the initial modules in the sequence create the desired pyrazole
core from readily available commercial building blocks. Once
the pyrazole core has been generated, subsequent modules
modify and augment the substituents appended to these cores
to generate the molecule of interest (Figure 2). The structure of
pyrazole product is controlled by the type and sequence of
modules installed in the assembly line and the starting materi-
als supplied to the system. In building upon our previous work,
module 1 rapidly creates diazoalkanes that are immediately
consumed in a [3+2] cycloaddition with an alkyne in module 2
to generate pyrazole cores. A range of additional modules then
modify this core, including module 3 that mediates C–N aryl-
ation, module 4 that mediates C–N methylation, module 5 that
mediates TMS-cleavage, and module 6 that mediates amidation
(Figure 2). In this work, orientation of the modules in the order:
1, 2, 5, 4, and then 6 enabled parallel syntheses of Fluxapyroxad
and Bixafen (Figure 2, Route C). While replacing modules 5, 4,
and 6 with an offline Ullmann coupling generated Celecoxib,
Mavacoxib, and SC-560 in a parallel fashion (Figure 2, Route D).
For convenience, each module is described individually below.
Figure 3. Optimizing the synthesis of 3 in modules 1 and 2. (A) A schematic
of the continuous flow set-up. General conditions: CF2HCH2NH2 (2.0 equiv.)
and AcOH (0.4 equiv.) diluted in CHCl3, tBuONO (2.4 equiv.) diluted in CHCl3,
and the alkyne (0.19 M, 1.0 equiv.) diluted in CHCl3. (B) Temperature optimiza-
tion of module 1. Module 1: tR = 5 min; module 2: tR = 30 min at 90 °C. (C)
Temperature optimization of module 2. Module 1: tR = 20 min at 60 °C; mod-
ule 2: tR = 30 min. (D) Residence time optimization of module 1. Module 1:
60 °C; module 2: tR = 30 min at 90 °C. (E) Residence time optimization of
module 2. Module 1: tR = 5 min at 60 °C; module 2: 90 °C. Optimized condi-
tions for the synthesis of 3: Module 1: tR = 5 min at 60 °C; Module 2: tR
=
10 min at 90 °C. Error bars indicate the standard deviation around the mean
(n = 2).
Using TMS-protected alkynes in module 2 affords pyrazoles
with electron withdrawing functionality at the 4-position
(Table 1 and Table 2, 5–8). However, applying the general con-
ditions above for the synthesis of 5 yielded only 44 % (Table 1,
Entry 1). Analysis of the reaction mixture led to the isolation of
both 5 and 3. While 5 is formed through [3+2] cycloaddition
between 1 and the TMS-alkyne, 3 is formed due to TMS cleav-
age by AcOH prior to cycloaddition, leading to the electroni-
cally favored isomer 3. To avoid TMS cleavage, the concentra-
Figure 2. An overview of the assembly line concept showing the six continu-
ous flow modules and one batch module. The initial concept focused on
route A to synthesize AS-136A. Here, route C is used to synthesize two agro-
chemicals, and route D is used to synthesize three pharmaceuticals.
(i) Module 1 and 2: Making the Pyrazole Core
In this assembly line process, common pyrazole cores are con- tion of AcOH was reduced to 10 %, and the alkyne operated in
structed through [3+2] cycloaddition between 2,2-difluoro- twofold excess relative to the amine. With these conditions, 5
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
2
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
was afforded in 59 % yield and 100 % selectivity, and 6 in 72 %
During optimization, we found that replacing fluoroamines
yield, and 76 % selectivity (Table 1, Entry 5). However, synthesiz- (CF2RCH2NH2) with Et3N, generated symmetrical divinyl ether 9
ing the CF3 analogue 7 with these new conditions lead to only in quantitative yield under identical conditions (Figure 4A). Sev-
2 % yield, while initial conditions yielded only 21 % (Table 2, eral control reactions revealed that this is likely the product of
Entry 1 and 2). Re-optimization revealed that a fourfold excess a radical chain process: First, reacting ethyl propiolate and
of both 2,2,2-trifluoroethylamine and tBuONO with 15 % AcOH tBuONO under identical conditions, but devoid of Et3N, yielded
was necessary for high yields and selectivity (Table 2, Entry 6). 0 %, and returned only starting material. Second, removing
With these conditions, 7 was afforded in 70 % yield and 85 % tBuONO from the reaction mixture yielded 0 %, revealing that
selectivity, and 8 in 67 % yield, and 88 % selectivity. This deli- both Et3N and tBuONO are required for the generation of 9.
cate optimization process revealed that when using substrates Next, adding stoichiometric TEMPO to the reaction mixture de-
containing CF3 or CF2H moieties, minor alterations may be re- creased the yield to 86 %, indicating a possible radical pathway.
quired to ensure both good yield and selectivity. Adding a new Finally, sparging solutions with argon prior to the reaction to
pyrazole core to the assembly line library simply requires vali- remove dissolved O2 led to 9 still being generated in quantita-
dating, and if necessary, re-optimizing, the conditions em- tive yields. This final result indicates that the vinyl oxygen atom
ployed in modules 1 and 2.
present in 9 most likely comes from tBuONO.
Table 1. Optimization studies for the synthesis of 5 over byproduct 3.
Entry
tBuONO/alkyne ratio
Amine/alkyne ratio
AcOH/alkyne
[%]
Yield 5[a]
[%]
Yield 3[a]
[%]
1
2
3
4
5
2.4
4.8
0.4
0.6
0.9
2.00
4.00
0.33
0.50
0.50
40
80
7
7
10
44
34
61
43
59
20
41
4
0
0
[a] Average isolated yield (n = 2). In this optimization, the concentration of AcOH and the ratio of reagents were varied for high selectivity and yield. For a
list of additional pyrazoles synthesized, see previous work.[6]
Table 2. Optimization studies for the synthesis of 7 over byproduct 4.
Entry
tBuONO/alkyne ratio
Amine/alkyne ratio
tR Module 1
[min]
tR Module 2
[min]
AcOH/alkyne
[%]
Yield 7[a]
[%]
Yield 4[a]
[%]
1
2
3
4
5
6
0.9
2.4
4.0
4.5
3.6
4.0
0.5
2.0
3.0
3.0
3.6
4.0
5.0
5.0
8.3
8.3
8.3
5.0
10.0
10.0
16.7
16.7
16.7
10.0
10
40
15
15
15
15
2
0
6
2
3
0
12
21
16
33
45
70
[a] Average isolated yield (n = 2). In this optimization, the concentration of AcOH and the ratio of reagents were varied for high selectivity and yield. For a
list of additional pyrazoles synthesized, see previous work.[6]
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
3
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 4. Synthesis of divinyl ether 9. (A) A schematic of the continuous flow reactor set-up. (B) Proposed mechanism: (i) Thermal bond homolysis of tBuONO
to generate tBuO radical. (ii) Addition of the tBuO radical into ethyl propiolate. (iii) Proton abstraction from pendant tBu group. (iv) Propagation of vinyl ether
radical into another molecule of ethyl propiolate. (v) Proton abstraction from Et3N. Optimized conditions: Et3N (0.193
M, 1 equiv.) and AcOH (0.077 M, 40 %)
in DCM, tBuONO (0.193 , 1 equiv.) in DCM, and ethyl propiolate (0.387 , 2 equiv.) in DCM. A crystal structure to aid in the assignment of 9 was obtained
M
M
by recrystallization in hexanes/EtOAc (10:1). The crystal structure has been added to the Cambridge Structural Database (#1572340).
To explain this unusual transformation, we propose the fol- AcOH maybe activating tBuONO towards bond homolysis. To
lowing mechanism. First, heating tBuONO generates a tBuO· test this hypothesis, the reaction temperature was reduced to
radical that adds across the alkyne generating a vinyl centered 40 °C (below the typical decomposition temperature of
radical (Figure 4B, step i).[12] After proton abstraction from the tBuONO[12b]), and 9 was still afforded in 95 % yield. At this
pendant tBu group, a terminal O· radical is generated that can stage, a radical pathway seems likely, but a two-electron mech-
propagate onto another molecule of ethyl propiolate (Figure 4B, anism cannot be ruled out before additional studies.
steps ii–iii). After propagation, Et3N behaves as a proton donor
(ii) Module 3: C–N Arylation
to afford 9 (Figure 4B, steps iv–v). Interestingly, this reaction
requires 40 % AcOH to achieve high yields as running the reac- Our initial communication briefly detailed a continuous flow
tion devoid of AcOH affords only 31 % yield. We believe that C–N arylation through a Cu(OAc)2 catalyzed Chan–Lam–Evans
Table 3. Optimizing C–N arylation in module 3.
Entry
Concentration
4 [
tR
[min]
PhB(OH)2
[equiv.]
Conditions[a]
Base
[equiv.]
Yield 10
[%]
M
]
1
2
3
4
5
6
7
8
0.16
0.32
0.16
0.16
0.32
0.32
0.32
60
60
60
60
60
90
60
60
1.50
1.50
1.50
1.50
0.50
0.50
0.50
0.50
microwave
microwave
microwave
microwave
microwave
microwave
flow
1.0
1.0
2.0
–
–
–
31
16
19
45/26[b]/34[c]
55
56
–
–
49
0.16
flow
66/42[d]
[a] Microwave conditions refers to batch microwave reactions (Biotage), not continuous flow microwave conditions. [b] Under 1 atm of O2. [c] TEMPO
(1.0 equiv.) was added. [d] Cu(OAc)2 (20 mol-%) and TEMPO (1.0 equiv.). General reaction conditions: tr = 60 min at 90 °C, 4 (0.12 , 2.0 equiv.), phenylboronic
acid (0.06
, 1.0 equiv.), Cu(OAc)2 (1.5 equiv.) and pyridine (2.0 equiv.). Yields were determined through 1H NMR analysis with anisole as an external standard
(26 μL, 0.240 mmol) after isolation by column chromatography.
M
M
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
4
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
coupling. For optimization of this reaction, high throughput
screening of reaction conditions was performed in a batch mi-
crowave reactor (Table 3). Screening reactions under microwave
conditions prior to translation into continuous flow has several
benefits. First, solid formation in a microwave vial is less likely
to cause problems; in continuous flow, solid formation can lead
to reactor failure. Second, commercial microwave reactors allow
reaction automation; a feature not currently readily available in
continuous flow reactors. Third, less material is used in a micro-
wave reaction, as microwave processes do not require equilibra-
tion. Finally, once optimized, translating into continuous flow
can have additional benefits including reduced solvent vapori-
zation and enhanced mixing.
Initially, reacting 4 with phenylboronic acid at 90 °C for
60 min with Et3N and pyridine afforded 10 in 31 % yield
(Table 3, Entry 1). Through optimization of reaction tempera-
ture, solvent, concentration of 4 and Et3N, we found that per-
forming the reaction devoid of Et3N and with a low concentra-
tion of 4 was vital (45 % yield, Table 3, Entry 4). Furthermore,
addition of an external oxidant such as TEMPO, or conducting
the reaction under an atmosphere of O2 had no benefit
(Table 3, Entry 4). With these conditions, the reaction was trans-
lated into continuous flow to obtain a 49 % yield (Table 3, Entry
7). To increase the yield of 10 to 66 %, it was crucial to use 4
in low concentrations, and in twofold excess relative to phenyl-
boronic acid (Table 3, Entry 8); we hypothesize that 4 is a com-
petitive ligand, and its use as the rate-limiting reagent will
always generate low yields. Finally, reducing the loading of
Cu(OAc)2 to 20 % and adding one equivalent of TEMPO af-
forded 10 in 42 % yield, indicating a catalytic approach may be
possible in future work.
Figure 6. Synthesis of SC-560 17, Mavacoxib 18, Celecoxib 19, and sulfon-
amide 20. (A) Batch Ullmann couplings for the synthesis of SC-560, Mava-
coxib, and Celecoxib. Optimal conditions: Pyrazole (0.50 mmol, 1.0 equiv.),
aryl iodide (4.0 equiv.), CuI (5 mol-%), N,N′-dimethylcyclohexyldiamine
(20 mol-%), K2CO3 (2.1 equiv.) and 1,4-dioxane (2 mL). Reactions were con-
ducted in 20 mL Biotage microwave vials at 150 °C for 24 h. (B) A schematic
of the continuous flow set-up used for the synthesis of sulfonamide 20.
Although Cu(OAc)2 catalyzed C–N arylation generated 10 in
good yield, these conditions were not successful for pyrazoles
11 and 12; <11 % yields were obtained in both cases (Figure 5).
We hypothesize that the increased sterics of 11 and 12 lead
to the requirement for longer reaction times and more forcing
conditions. Indeed, recent reports require heating the reaction
mixture at 150 °C for 36 h with catalytic CuI (5 mol-%), K2CO3,
and diamine ligand (20 mol-%).[13] Thus, to furnish SC-560, Ma-
vacoxib, and Celecoxib in good yield, batch Ullman couplings
were performed. For the synthesis of SC-560 17, pyrazole 15
was first synthesized in modules 1 and 2 (76 %, 32 min)[6] before
an Ullmann coupling with 4-iodoanisole (24 h, 150 °C). Under
these conditions, the human lung, breast, and colon cancer sup-
pression candidate 17 was afforded in 70 % yield (Figure 6A).[10]
For the synthesis of Mavacoxib 18 and Celecoxib 19, sulfon-
amide 20 is required as the coupling partner. To aid in the
synthesis of both Celecoxib and Mavacoxib, the synthesis of 20
was translated to a continuous flow system. Here, reacting 4-
iodobenzenesulfanoylchloride with dibenzylamine for 6.7 min
generated 20 in quantitative yield at 9.3 g h–1 (Figure 6B). Addi-
tionally, an in line quench and extraction allowed purified 20
to be generated in 73 % yield at 6.8 g h–1. Next, pyrazoles 15
and 16 were synthesized in modules 1 and 2 (76 % yield,
32 min)[6] before Ullmann couplings with 20 (24 h, 150 °C) fur-
nished the protected forms of Mavacoxib and Celecoxib. To re-
veal the bioactive compounds, deprotection with conc. H2SO4
afforded Mavacoxib 18 and Celecoxib 19 in 43 % and 71 %
yield respectively (Figure 6B). In concluding this module, com-
bining continuous flow and batch chemistry allows efficient ac-
cess to several bioactive molecules in a parallel fashion. Here,
Ullmann couplings were performed in batch due to their long
reaction times (24 h). Development of a more continuous flow
friendly process would be highly desirable.
(iii) Module 4: C–N Methylation
Several of the APIs and agrochemicals require C–N methylation
at the N1 position. In previous syntheses, methylation is
achieved by using iodomethane (MeI) and K2CO3.[14] The use of
heterogeneous reaction mixtures in continuous flow, however,
Figure 5. C–N arylated products using the optimized conditions described in
Table 3. To note is that pyrazoles 11, 12, 15, and 16 are synthesized via
modules 1 and 2 as per our previous report.[6]
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
5
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
is not ideal. To overcome this, several organic bases were reaction conditions for both CF2H, and CF3 pyrazoles, DBU was
screened for the methylation of 4 under microwave conditions. found to yield > 99 % and 85 % respectively (Table 4, Entries 9
Of the bases tested (Et3N, DABCO, Hünig's base) tetramethyl- and 10). Translating optimized conditions from microwave ex-
piperidine (TMP) at 120 °C for 30 min afforded 21 in 99 % yield periments to a single continuous flow set up with DBU pro-
(Table 4, Entry 2). Additionally, reducing the reaction time to vided satisfactory conditions for methylating both CF2H and CF3
one min under the same conditions afforded 21 in 80 % yield analogues in 94 % and 79 % yield respectively (Table 4, Entry
(Table 4, Entry 3), while at 22 °C only 64 % was obtained 11 and 12). Lastly, translating into the multi-step continuous
(Table 4, Entry 4). Next, several mixtures of DMF and DCM were flow system (Table 4, Scheme B) with an extended reaction time
tested to simulate the conditions expected during reaction tele- of 27.6 min and increasing to five equivalents of MeI and DBU,
scoping with modules 1 and 2. While a 75:25 mixture of DMF/ 21 and 22 were yielded in 81 % and 75 % respectively (Table 4,
DCM for one min at 90 °C afforded 21 in 65 % yield (Table 4, Entries 13 and 14).
Entry 5), pure DCM under the same conditions afforded only
6 % yield (Table 4, Entry 8).
Another benefit of this multi-step system was the ability to
avoid MeI decomposition in DMF. Pre-diluting MeI with DMF
Next, CF2H pyrazole 3 was tested, but only a 29 % yield of leads to slow generation (> 30 min) of a yellow solution; MeI
22 was obtained. We hypothesize that the acidity of the CF2H reacts with DBU releasing iodide. To avoid this, neat MeI and
proton leads to additional and unwanted reactivity. In exploring DMF from the supplier's bottle were diluted in situ to create the
Table 4. Optimization of C–N methylation.
Entry
R
Reactor temp.
[°C]
tR
[min]
MeI
[equiv.]
Conditions
Base
Base
[equiv.]
DMF/DCM ratio
Yield
[%]
1
2
3
4
5
6
7
8
F
F
F
F
F
F
F
F
H
F
H
F
H
F
120
120
120
22
90
90
30
30
1
1
1
1
1
1
30
30
10
10
27.6
27.6
1.3
2.0
2.0
2.0
2.0
2.0
2.0
2.0
4.0
4.0
4.0
5.0
5.0
5.0
microwave
microwave
microwave
microwave
microwave
microwave
microwave
microwave
microwave
microwave
single-flow
single-flow
multi-flow
multi-flow
TMP
TMP
TMP
TMP
TMP
TMP
TMP
TMP
DBU
DBU
DBU
DBU
DBU
DBU
1.3
2.0
2.0
2.0
2.0
2.0
2.0
2.0
4.0
4.0
4.0
5.0
5.0
5.0
100
100
100
100
75:25
50:50
25:75
0
100
100
69
99
80
64
65
60
43
6
90
90
9
120
120
120
120
120
120
> 99[a,b]
85[a]
94[a,c]
79[a]
75 (77)[a]
81 (93)[a]
10
11
12
13
14
100
100
66:34
66:34
[a] Pyrazole concentration = 0.264 M
. [b] Methylation of 3 yielded 79:21 of the N1 methylated: N2 methylated regioisomer. [c] Methylation of 3 yielded 62:38
of the N1 methylated: N2 methylated regioisomer. Microwave reactions were performed with 0.19 mmol of 4, and yields were determined through 1H NMR
analysis with anisole as an external standard (0.10 mmol, 10.8 μL). For single and multi-step experiments, isolated yields are reported. Conversions are detailed
in parentheses.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
6
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 7. Synthesis of regioisomers 22A and 22B. Optimal reaction conditions: 23 (0.263
M
in DMF, 1.0 equiv.), MeI (1.05
M
in DMF, 4.0 equiv.), DBU (1.05
M
in
DMF, 4.0 equiv.). Regioisomers were separated by column chromatography (7:3 hexanes/EtOAc): Rf 22A = 0.25 and Rf 22B = 0.65.
correct concentration prior to entering the system. To further
highlight this point, premixing MeI and DBU in DMF and load-
ing into an 8 mL stainless steel syringe for elution over a 8 h
period at 12.5 μL min–1 leads to no methylation of 3, and re-
turns only starting material.
Table 5. Optimizing TMS cleavage in module 5.
When methylating CF3 bearing pyrazoles, only the N1 meth-
ylated pyrazole was observed. However, when methylating
CF2H containing pyrazole 3, a ≈ 8:2 mixture of N1 and N2 meth-
ylated regioisomers was obtained. With the methylation of di-
fluoro analogue 22 required for the synthesis of Fluxapyroxad
and Bixafen (Figure 1), we found conditions to separate the
regioisomers by column chromatography on silica. In this case,
methylating 22 yields a 62:38 mixture of N1 and N2 regioiso-
mers (22A and 22B, Figure 7).
Entry
R
Nucleophile Conditions[a]
[equiv.]
Yield
[%][a]
1
2
3
4
5
H
H
H
F
2.44
2.44
1.75
1.22
1.75
TBAF with AcOH (2.44 equiv.), DCM
TBAF with AcOH (2.44 equiv.), THF
TBAF with AcOH (1.75 equiv.), THF 96
TBAF with AcOH (1.22 equiv.), DCM 45
TBAF with AcOH (1.75 equiv.), THF 84[b]
62
74
F
[a] Yields were determined through 1H analysis with anisole (5.8 μL,
0.05 mmol). [b] Yield determined through 1H analysis with α,α,α-trifluoro-
toluene (30.0 μL, 0.246 mmol).
(iv) Module 5: TMS Removal
To afford ester functionality at the 4-position of the pyrazole
core (5–8), TMS-protected alkynes are used to alter the regio-
chemical outcome of the cycloaddition in module 2. For the
final targets, however, TMS cleavage is required. This module
required fine optimization to achieve TMS removal in the pres-
ence of CF2H groups, again highlighting the different condi-
tions required for CF3 and CF2H functionalities. First, TMS cleav-
age of 8 was explored under microwave conditions. Inspired
by previous reports[15] heating at 150 °C for 30 min with two
equivalents of KOH yielded 24 in 68 % yield. However, when
these conditions were tested for TMS cleavage of 6, 0 % yield
resulted. TLC analysis of the crude reaction mixture revealed full
(v) Module 6: Amidation
For many targets, amidation reveals the final bioactive struc-
tures. First, direct amidation of the pyrazole ester was at-
tempted using high temperature microwave conditions. Using
DMF as the solvent, a 2:1 ratio of 4-fluoroaniline: 21 at 170 °C
for 30 min resulted in no reactivity. To overcome this, our previ-
ous approach using LiHMDS was used.[6,16] Given the eventual
telescoped nature of this process, the solvent was switched
from DCM to DMF; an excellent 94 % yield was still obtained
(Table 6, Entry 2). Next, while reducing the residence time to
2.5 min gave quantitative yields, a one min residence time af-
forded 83 % (Table 6, Entries 3–4). To overcome this drop in
yield, the reactor coil was heated to 40 °C to afford 35 in 93 %
1
consumption of starting material, while H NMR indicated that
no CF2H group was present. In exploring new deprotection
strategies, the use of 2.4 equivalents of tetrabutylammonium
fluoride (TBAF) buffered with equimolar AcOH was suitable. Un-
der these conditions, 24 was afforded in 74 % yield (Table 5,
Entry 2), while lowering the concentration of TBAF and AcOH
to 1.75 equivalents lead to 96 % (Table 5, Entry 3). Furthermore,
when using CF2H analogue 6, an 84 % yield of 23 was obtained
(Table 5, Entry 5). To note, not buffering with AcOH lead to
decomposition of 6, and testing fluoride on polymer support
isolated yield (Table 6, Entry 5). Notably, the 1
M LiHMDS in THF
was used directly from the supplier's bottle by interfacing the
inlet feed of the stainless steel syringe to a needle, which is
then attached to commercial LiHMDS using standard syringe
techniques. This generalizable flow approach avoids user expo-
sure to this flammable and toxic chemical.
With amidation optimized, the synthesis of agrochemicals
(Sigma, 20–50 mesh, 2–3 mmol/g loading) showed no reactivity Fluxapyroxad and Bixafen was achieved. First, the amines re-
with 6.
quired for generating Fluxapyroxad and Bixafen were synthe-
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
7
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 8. Continuous flow synthesis of Fluxapyroxad and Bixafen along with their respective N2 regioisomers. (A) Synthesis of the required amines. (B)
Continuous flow synthesis of the shown agrochemicals. Conditions used: Pyrazole (0.30
M in DMF, 1.00 equiv.), amine (1.33 equiv. in DMF), and 1 M LiHMDS
(3.33 equiv. in THF) were reacted in a 250 μL reactor at 40 °C. Solutions were collected and quenched upon exit by NH4Cl.
Table 6. Optimizing amidation in module 6.
in 90 % and 99 % respectively (Figure 8B). In conclusion, Fluxa-
pyroxad and Bixafen have been synthesized in 38 % overall
yield through using five different continuous flow modules (two
of which are telescoped) for an overall reaction time of only
56 min. To further extend the utility of this process, their N2
regioisomers were also synthesized to demonstrate the parallel
nature of this assembly line synthesis.
(vi) Enzymatic Hydrolysis of Pyrazole Ester 4
Pyrazole ester hydrolysis is currently the major pathway to gen-
erate the required amides. Typically, 1
ester, 6
M NaOH hydrolyzes the
M
HCl generates the free acid, and SOCl2 activates the
acid for amidation.[13b,17] With interests in developing additional
methods for manipulating pyrazoles, enzymatic hydrolysis with
an immobilized lipase was explored (Figure 9). Dissolving 4 in
Entry
Reactor temp.
[°C]
tR Module 5
[min]
Pyrazole solvent
Yield
[%]
1
2
3
4
5
23
23
23
23
40
5.0
5.0
2.5
1.0
1.0
DCM
DMF
DMF
DMF
DMF
91
94
100
83
93
sized through a multi-gram Suzuki cross couplings in batch;
both proceeded in > 90 % yield (26 and 27, Figure 8A). Next,
N1 methylated pyrazole 22A was treated with both 26 and 27
to afford Fluxapyroxad 28 and Bixafen 29 in 96 % yield (Fig-
ure 8B). Then, the N2 regioisomer was passed through the sys-
tem to afford regioisomers of Fluxapyroxad 30 and Bixafen 31
Figure 9. Enzymatic hydrolysis of 4 to yield free acid 32 with immobilized
CAL-B lipase.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
8
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
3449–3464; Angew. Chem. 2015, 127, 3514–3530; d) S. V. Ley, D. E. Fitz-
patrick, R. M. Myers, C. Battilocchio, R. J. Ingham, Angew. Chem. Int. Ed.
2015, 54, 10122–10136; Angew. Chem. 2015, 127, 10260–10275; e) M.
Movsisyan, E. I. P. Delbeke, J. K. E. T. Berton, C. Battilocchio, S. V. Ley, C. V.
Stevens, Chem. Soc. Rev. 2016, 45, 4892–4928; f) D. Webb, T. F. Jamison,
Chem. Sci. 2010, 1, 675–680; g) C. Wiles, P. Watts, Green Chem. 2012, 14,
38–54.
DMSO and diluting with PBS buffer at pH 7 provided a homoge-
neous solution (96 μmol in 99:1 H2O/DMSO) suitable for hydrol-
ysis with immobilized CAL-B lipase (Candida antarctica recombi-
nant from Aspergillus oryzae). Simply stirring the solution for 8 h
at 40 °C, and filtering off the immobilized lipase provided 32
in 75 % yield. In demonstrating the utility of this reaction, the
transformation was performed on an 800 mg scale; 32 was af-
forded in 37 % isolated yield. This simple enzymatic hydrolysis
adds a useful tool of this assembly line sequence that impor-
tantly removes the need for concentrated HCL and NaOH.
[2] a) T. Razzaq, C. O. Kappe, Chem. Asian J. 2010, 5, 1274–1289; b) J. We-
gner, S. Ceylan, A. Kirschning, Adv. Synth. Catal. 2012, 354, 17–57; c) T. N.
Glasnov, C. O. Kappe, Chem. Eur. J. 2011, 17, 11956–11968.
[3] a) B. J. Deadman, S. G. Collins, A. R. Maguire, Chem. Eur. J. 2015, 21,
2298–2308; b) B. Gutmann, D. Cantillo, C. O. Kappe, Angew. Chem. Int.
Ed. 2015, 54, 6688–6728; Angew. Chem. 2015, 127, 6788–6832; c) S. T. R.
Müller, T. Wirth, ChemSusChem 2015, 8, 245–250.
[4] a) K. Jähnisch, V. Hessel, H. Löwe, M. Baerns, Angew. Chem. Int. Ed. 2004,
43, 406–446; Angew. Chem. 2004, 116, 410–451; b) Y. Su, K. Kuijpers, V.
Hessel, T. Noel, React. Chem. Eng. 2016, 1, 73–81; c) K. P. L. Kuijpers,
M. A. H. van Dijk, Q. G. Rumeur, V. Hessel, Y. Su, T. Noel, React. Chem.
Eng. 2017, 2, 109–115.
[5] a) D. Ghislieri, K. Gilmore, P. H. Seeberger, Angew. Chem. Int. Ed. 2015,
54, 678–682; Angew. Chem. 2015, 127, 688–692; b) J. Britton, J. W. Castle,
G. A. Weiss, C. L. Raston, Chem. Eur. J. 2016, 22, 10773–10776; c) J. Brit-
ton, J. M. Chalker, C. L. Raston, Chem. Eur. J. 2015, 21, 10660–10665.
[6] J. Britton, T. F. Jamison, Angew. Chem. Int. Ed. 2017, 56, 8823–8827;
Angew. Chem. 2017, 129, 8949–8953.
Conclusions
Using the concept of an assembly line, five bioactive molecules
have been synthesized in a parallel fashion. Using six continu-
ous flow, and one batch reactor, low cost fluorinated amines
are transformed into agrochemicals and pharmaceuticals in
short order. The success of this system stems from the ability of
the continuous flow modules to operate solvents and reagents
above their atmospheric boiling points, and safely process toxic
and explosive intermediates. Using five different continuous
flow modules (four individual continuous flow processes),
Bixafen and Fluxapyroxad were synthesized in a 38 % overall
yield. Additionally, coupling a telescoped continuous flow proc-
ess with an Ullmann coupling in batch provided access to SC-
560, Celecoxib, and Mavacoxib in 43–70 % yield over 2–3 steps.
The ability to synthesize many bioactive molecules by modify-
ing a common core opens up the possibility for generating
compound libraries for medicinal chemistry. Additionally, as
many natural products contain the same core functionality, we
envisage the ability to create families of natural products
through a similar assembly line process.
[7] F. Giornal, S. Pazenok, L. Rodefeld, N. Lui, J.-P. Vors, F. R. Leroux, J. Fluorine
Chem. 2013, 152, 2–11.
[8] Ş. G. Küçükgüzel, S. Şenkardeş, Eur. J. Med. Chem. 2015, 97, 786–815.
[9] a) L. Y. Pang, S. A. Argyle, A. Kamida, K. O. N. Morrison, D. J. Argyle, BMC
Vet. Res. 2014, 10, 184; b) S. R. Cox, S. Liao, M. Payne-Johnson, R. J.
Zielinski, M. R. Stegemann, J. Vet. Pharmacol. Ther. 2011, 34, 1–11.
[10] a) E. Lee, M.-K. Choi, H.-J. Youk, C. H. Kim, I.-O. Han, B.-C. Yoo, M.-K. Lee,
S.-J. Lim, J. Cancer Res. Clin. 2006, 132, 223–233; b) S. Grösch, I. Tegeder,
E. Niederberger, L. Bräutigam, G. Geisslinger, FASEB J. 2001, 15, 2742–
2744; c) N. Kundu, M. J. Smyth, L. Samsel, A. M. Fulton, Breast Cancer Res.
Treat. 2002, 76, 57–64.
[11] a) K. Monir, M. Ghosh, S. Jana, P. Mondal, A. Majee, A. Hajra, Org. Biomol.
Chem. 2015, 13, 8717–8722; b) S. Maity, T. Naveen, U. Sharma, D. Maiti,
Org. Lett. 2013, 15, 3384–3387; c) Y.-f. Ji, H. Yan, Q.-b. Jiang, Eur. J. Org.
Chem. 2015, 2051–2060; d) C.-X. Miao, B. Yu, L.-N. He, Green Chem. 2011,
13, 541–544.
CCDC 354833 (for 9) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge from
The Cambridge Crystallographic Data Centre.
[12] a) C. Tang, T. Shen, N. Jiao, Nitrogenation Strategy for the Synthesis of N-
containing Compounds, Springer Nature Singapore, 2017; b) U. Dutta,
D. W. Lupton, D. Maiti, Org. Lett. 2016, 18, 860–863.
[13] a) Y. Wang, J. Han, J. Chen, W. Cao, Tetrahedron 2015, 71, 8256–8262; b)
F. Li, J. Nie, L. Sun, Y. Zheng, J.-A. Ma, Angew. Chem. Int. Ed. 2013, 52,
6255; Angew. Chem. 2013, 125, 6375–6378.
Supporting Information (see footnote on the first page of this
article): Experimental procedures and synthesis.
[14] L. Mertens, K. J. Hock, R. M. Koenigs, Chem. Eur. J. 2016, 22, 9542–9545.
[15] J. E. Curry, J. F. Lyons, M. S. Squires, N. T. Thompson, M. M. Thompson,
P. G. Wyatt, Vol. WO 2006077425 (Ed. W. I. P. Organization), Astex Thera-
peutics Limited, UK, USA, 2006.
Acknowledgments
We thank the Bill and Melinda Gates Foundation (“Medicines
for All” initiative) for financial support.
[16] J. L. Vrijdag, F. Delgado, N. Alonso, W. M. De Borggraeve, N. Perez-Macias,
J. Alcazar, Chem. Commun. 2014, 50, 15094–15097.
[17] a) B. Pelcman, A. Sanin, P. Nilsson, K. No, W. Schaal, S. Öhrman, C. Krog-
Jensen, P. Forsell, A. Hallberg, M. Larhed, T. Boesen, H. Kromann, S. B.
Vogensen, T. Groth, H.-E. Claesson, Bioorg. Med. Chem. Lett. 2015, 25,
3024–3029; b) H. Jasch, J. Scheumann, M. R. Heinrich, J. Org. Chem. 2012,
77, 10699–10706; c) L. Xiong, H. Li, L.-N. Jiang, J.-M. Ge, W.-C. Yang, X. L.
Zhu, G.-F. Yang, J. Agric. Food Chem. 2017, 65, 1021–1029; d) X.-H. Liu,
W. Zhao, Z.-H. Shen, J.-H. Xing, T.-M. Xu, W.-L. Peng, Eur. J. Med. Chem.
2017, 125, 881–889.
Keywords: Continuous flow · Assembly line synthesis ·
Synthetic methods · Diazo compounds · Nitrogen
heterocycles
[1] a) J. Britton, C. L. Raston, Chem. Soc. Rev. 2017, 46, 1250–1271; b) R. L.
Hartman, K. F. Jensen, Lab Chip 2009, 9, 2495–2507; c) S. V. Ley, D. E.
Fitzpatrick, R. J. Ingham, R. M. Myers, Angew. Chem. Int. Ed. 2015, 54,
Received: July 14, 2017
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
9
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Continuous Flow Synthesis
J. Britton, T. F. Jamison* ............... 1–10
Synthesis of Celecoxib, Mavacoxib,
SC-560, Fluxapyroxad, and Bixafen
Enabled by Continuous Flow Reac-
tion Modules
Azoles Assemble! Multi-step continu- and pharmaceuticals Mavacoxib, Cele-
ous flow enables the synthesis of agro- coxib, and SC-560.
chemicals Fluxapyroxad and Bixafen,
DOI: 10.1002/ejoc.201700992
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
10
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim