Development of a Safe and Robust Process for the Large-Scale Preparation of a Vinyl Bromide from a Ketone Using a (PhO)3P/Br2-Derived Reagent

Article abstract of DOI:10.1021/acs.oprd.6b00100

The large-scale synthesis of ethyl 4-bromocyclohex-3-enecarboxyalate, using a mild brominating reagent derived from triphenyl phosphite and bromine, is reported. The development and comparison of both continuous and batch processes are described. A modifi

Full text of DOI:10.1021/acs.oprd.6b00100

Article
pubs.acs.org/OPRD
Development of a Safe and Robust Process for the Large-Scale
Preparation of a Vinyl Bromide from a Ketone Using
a (PhO)3P/Br2‑Derived Reagent
Nachiket Likhite,† Sivaraj Ramasamy,† Shankar Tendulkar,† Shunmugaraj Sathasivam,† Michael Luzung,‡
Ye Zhu,‡ Neil Strotman,‡ Jeffrey Nye,‡ Adrian Ortiz,‡ Susanne Kiau,‡ Martin D. Eastgate,‡
and Rajappa Vaidyanathan*,†
†Chemical and Synthetic Development, Biocon Bristol-Myers Squibb Research and Development Center, Biocon Park, Jigani Link Road,
Bommasandra IV, Bangalore-560099, India
‡Chemical and Synthetic Development, Bristol-Myers Squibb, 1 Squibb Drive, New Brunswick, New Jersey 08903, United States
ABSTRACT: The large-scale synthesis of ethyl 4-bromocyclohex-3-enecarboxyalate, using a mild brominating reagent derived
from triphenyl phosphite and bromine, is reported. The development and comparison of both continuous and batch processes
are described. A modified addition sequence was developed based on the knowledge garnered from flow-processing, resulting in a
safe and efficient process for the in situ generation of the unstable active reagent and its immediate reaction with the ketone in a
batch mode process.
learned during the course of developing the flow protocol paved
the way for the development of an efficient second-generation
batch process, vide infra.
INTRODUCTION
■
The increasing structural complexity of active pharmaceutical
ingredients and intermediates presents unprecedented chal-
lenges during their synthesis, especially on a large scale. Metal-
mediated cross-coupling reactions have emerged as powerful
tools in the synthetic chemist’s arsenal, enabling the rapid and
selective formation of complex structural motifs in a relatively
straightforward manner.1 As these technologies (i.e., Suzuki,
Suzuki−Miyaura, Buchwald−Hartwig, Sonogashira, Heck cou-
pling, etc.) continue to evolve, so too must the chemistry to
prepare the desired electrophilic cross coupling partners, such
as enol triflates and vinyl halides. While both of these electro-
philes are competent in cross-coupling transformations, the vinyl
bromides are generally far superior with respect to chemical
stabilitya key attribute often required in large-scale multistep
syntheses.
Traditional methods for the direct preparation of vinyl halides
from carbonyl compounds often involve the use of halogenating
agents under harsh conditions.2 Multistep processes have also
been reported involving the isolation of synthetic intermediates
which require further modification.3 Recently, Prati and co-workers
reported a mild protocol for the synthesis of vinyl halides
and gem-dihalides using triphenyl phosphite/halogen-based
reagents.4 This method exhibits good functional group compat-
ibility and is amenable to the synthesis of structurally diverse
halides from their corresponding carbonyl precursors. To the
best of our knowledge, there are no reports documenting the
application of this method for the large-scale synthesis of vinyl
halides.
RESULTS AND DISCUSSION
■
Vinyl bromide 5 (Scheme 1) is a key building block for the
synthesis of a new drug candidate currently under evaluation.
We envisioned that 5 could be synthesized from ketone 3 using
an appropriate brominating reagent. Based on literature prece-
dent, we explored the possibility of effecting this transformation
using Prati’s triphenyl phosphite-bromine-derived reagent,
presumed to be active in its ionic form 2a (Scheme 1).4
The first step in the original protocol was to prepare the
active brominating reagent (2a) by the addition of molecular
bromine to a solution of triphenyl phosphite 1 in dichloro-
methane, maintaining the temperature below −60 °C. Further
addition of triethylamine and starting material (ethyl cyclo-
hexanone-4-carboxylate, 3) led to the formation of the gem-
dibromo intermediate 4. Subsequent warming of the reaction
mixture to ambient temperature and elimination of HBr fur-
nished 5 upon workup. Our preliminary experiments suggested
that the reaction mixture had to be warmed to ambient
temperature and maintained for 24 h in order to effect the
desired elimination reaction (Scheme 1, 4 → 5).5
This first-generation batch mode process was adopted for
supporting early scale-up activities and delivered ∼70% yield of
5 (Figure 1).6 However, we encountered several challenges
during the execution of this process in the pilot plant, especially
while charging bromine, triethylamine, and starting material
within the stipulated times, controlling the reaction temperature
below −60 °C. Addition of molecular bromine to a solution of
triphenyl phosphite in dichloromethane was exothermic, such that
In this report, we describe the successful development and
application of the Prati method for the large-scale synthesis of
an important vinyl bromide intermediate. An initial batch
process highlighted some issues with the protocol, especially
around stability of the active reagent; thus a continuous flow
process was developed for reagent preparation. The lessons
Received: March 23, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.6b00100
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1
5 (Scheme 2). The ionic species 2a is reported to convert to
the inactive covalent form, dibromotriphenoxyphosphorane
[(PhO)3PBr2, 2b] over time even at low temperatures
(Scheme 2), and the ratio of the two species appears to be
temperature-dependent (2a:2b is reported to be 4:1 at −60 °C,
and 1:3 at −20 °C).4,7 Thus, in order for the bromination
process to be efficient, the bromination reaction should be faster
than the conversion of 2a to 2b.
Preliminary investigations revealed that the capricious perfor-
mance of the bromination reaction on scale may be attributed to
the long addition times in the pilot plant. We presume that the
root cause of this loss in reactivity is the depletion of the reactive
ionic form, 2a. A series of experiments was conducted to gain
further insights into the nature and stability of the brominating
reagent and its reactivity with 3. In one experiment, 2a was
generated and maintained at −60 °C for 16 h prior to the
addition of triethylamine and 3. This reaction, upon warming to
ambient temperature, provided 5 in ca. 75% yield. However, in
instances where 2a was generated at −60 °C and was allowed to
warm to −50 or −35 °C and held at those temperatures for 16 h
prior to the addition of triethylamine and 3, a much lower
yield (∼45%) was obtained upon workup, suggesting deterio-
ration in the reactivity of the system presumably by means of
conversion of 2a into the inactive form 2b above −50 °C,
consistent with literature reports.4 While these experiments are
consistent with prior observations, they do not discount a loss of
activity via other pathways.
Analysis of the mechanism of reagent formation and its
reaction with 3 led to the hypothesis that there is both a time and
temperature variable associated with the stability of the active
brominating agent. This led us to investigate the instantaneous
reaction of the brominating reagent 2a with the substrate 3, in
hopes of circumventing the possible loss of reagent quality
associated with on-scale processing times between reagent
formation and substrate addition. The goal of this work was to
develop a fundamentally robust process, unencumbered by tight
Figure 1. First-generation batch process.
the addition rate needed to be controlled carefully to maintain
the reaction temperature below −60 °C; for example, one plant-
scale batch (17 kg input) required an 8 h addition sequence
followed by an additional 8−9 h for warming the reaction mass
to ambient temperature, and resulted in a significantly lower
yield (55%) of the desired bromide 5. This result prompted us to
closely examine the reaction pathway and the unit operations in
the process in an effort to develop a more robust and efficient
process for the synthesis of 5.
In this reaction, it is believed that the reaction of bromine
and triphenyl phosphite leads to the active brominating agent,
bromotriphenoxyphosphonium bromide [(PhO)3P+BrBr−, 2a]
(Scheme 2).4,7 According to the literature, this active, ionic
species reacts with the substrate (3 in this case) to form the
oxyphosphonium bromide 6. Elimination of triphenylphosphate
from 6 provides the intermediate gem-dibromide 4, which
upon base-promoted dehydrobromination affords vinyl bromide
Scheme 2
B
DOI: 10.1021/acs.oprd.6b00100
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 2. First-generation semibatch process.
Figure 3. Second-generation semibatch process.
processing requirements (such as rigid control over processing
times). Since this appeared challenging in batch mode, we eval-
uated the possibility of generating the active reagent 2a under a
continuous process prior to charging the active stream into a
batch reactor containing substrate 3 and triethylamine (Figure 2).
Accordingly, precooled solutions of bromine in dichloromethane
(2.1 M, flow rate = 29 mL/min) and triphenyl phosphite 1 in
dichloromethane (2.1 M, flow rate = 45 mL/min)8 were pumped
and circulated through a 100 mL flow reactor (Hastelloy coil),
which was maintained at −50 to −60 °C (residence time =
1.35 min) before being charged into a batch reactor containing a
solution of 3 and triethylamine in dichloromethane at −5 °C.
This mixture was warmed to ambient temperature, and afforded
5 upon workup. This first-generation semibatch process was
found to be significantly more efficient than the original batch
process, resulting in ∼85% isolated yield of the desired product 5.
Interestingly, a sample analyzed immediately after completion
of addition of 2a showed almost complete consumption of ketone
3, even after minimal aging of the reaction stream post bromine
addition. This indicated that the reaction of 2a with 3 was
kinetically fast. Thus, we explored the possibility of generating
the active brominating agent 2a in the presence of ketone 3.
A second-generation semibatch process was therefore developed,
wherein two dichloromethane streamsone containing bro-
mine and the other a mixture of 3, triphenyl phosphite, and
triethylaminewere circulated through a flow reactor in similar
manner at −50 to −60 °C (bath temperature) and charged into
a batch reactor for further reaction and workup (Figure 3).
This second-generation semibatch process was also found to be
more efficient than the first-generation batch process and led
to ∼85% isolated yield of product 5.
The flow chemistry protocols enabled us to successfully generate
reagent 2a at −50 to −60 °C and treat it with ketone 3 instan-
taneously, resulting in both faster reactions and improved yields.
With this new knowledge, we sought to develop a more
robust and efficient second-generation batch process in an
effort to increase throughput. We postulated that charging
molecular bromine to a batch reactor containing a mixture of
3, triphenyl phosphite, and triethylamine in dichloromethane
at −40 to −50 °C may result in a reactivity and stability profile
suitable for reproducibly forming the desired bromide in a safe
and efficient manner. Gratifyingly, this subtle modification
gave a further improvement in yield and quality, providing 5 in
∼90% yield. After further parameter screening, we were able to
C
DOI: 10.1021/acs.oprd.6b00100
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
increase the reaction temperature and therefore significantly
improve reaction rate (time) and develop a second-generation
batch process, wherein molecular bromine was charged at
−20 to −5 °C to a mixture of 3, triphenyl phosphite, and
triethylamine in dichloromethane. This reaction proceeded
to completion within 1 h at 20 °C and generated vinyl bromide
5 in 90% yield (Figure 4).9 This modified second-generation
(14.2 kg, 139.8 mol) was added while maintaining the mass
temperature at −65 to −75 °C. Compound 3 (17 kg, 99.8 mol)
was charged to the reactor while maintaining the mass
temperature at −65 to −75 °C. The reaction mass was
maintained at −65 to −75 °C for 1 h and warmed to ambient
temperature over a period of 10 h and stirred for 23 h.
The reaction was quenched by charging 0.5 N HCl (85 L).
The layers were separated, and the organic layer was washed
with water (85 L). The organic layer was passed through a silica
gel bed (60−120 mesh size, 51 kg), and the bed was washed
with dichloromethane (87 L). The combined filtrate was
concentrated under vacuum to provide 5 (59.2 kg, assay by
HPLC: 21.7% w/w) as a viscous liquid in 55% yield.6
First-Generation Semibatch Process (Figure 2). Re-
agent solution A: a solution of bromine (25 mL, 485 mmol) in
dichloromethane (230 mL). Reagent solution B: a solution of
triphenyl phosphite (165 g, 531 mmol) in dichloromethane
(250 mL).
Reagent solution A was pumped at a rate of 29 mL/min
through one head of a double-headed Ceram pump Q2 V with
two Q1CTC pump heads and one V300 Stroke controller
(Make: Fluid Metering, Inc.). Reagent solution B was pumped
through the other head at 45 mL/min.8 The reagent solutions
A and B were flowed through 100 mL precooled tube reactors
maintained at −50 to −60 °C. These precooled solution streams
were mixed through a T-joint (static mixer). The mixed stream
was passed through a 100 mL coil maintained at −50 to −60 °C.
The outlet of the coil was drained to a reactor containing 3 (50 g,
294 mmol), triethylamine (75 mL, 538 mmol) in dichloro-
methane (500 mL) and maintained at −5 to 0 °C. Upon comple-
tion of the transfer, the mass was warmed to ambient temperature
and stirred for 2 h. Upon reaction completion, the reaction
mixture was quenched with 0.5 N HCl (500 mL), the layers were
separated, and the organic layer was washed with water (500 mL).
The organic layer was filtered through a silica gel pad (60−120
mesh size, 150 g), and the filtrate was concentrated under vacuum
to provide 5 (206 g, assay by HPLC: 28% w/w) as a viscous liquid
in 85% yield.6
Second-Generation Batch Process (Figure 4). A solution
of 3 (500 g 2.94 mol), triethylamine (580 mL, 4.16 mol), and
triphenyl phosphite (1100 g, 3.50 mol) in dichloromethane
(10 L) was stirred at −20 to −15 °C. To the solution was
charged bromine (180 mL, 3.49 mol) in a dropwise manner
while maintaining the mass temperature below −5 °C. Upon
completion of the addition, the mixture was warmed to ambient
temperature and stirred for 1 h. Upon reaction completion, the
reaction mixture was quenched with 0.5 N HCl (2.5 L), the
layers were separated, and the organic layer was washed with
water (2.5 L). The organic layer was filtered through a silica gel
pad (60−120 mesh size, 900 g) and the filtrate was concentrated
under vacuum to afford 5 (1950 g, assay by HPLC: 32% w/w)
as a viscous liquid in 90% yield.6
Figure 4. Second-generation batch process.
batch process was found to be more efficient than the first-gener-
ation batch process in many aspects as evident from Table 1.
CONCLUSION
■
In summary, we have successfully implemented the use of a mild
brominating reagent 2a derived from triphenyl phosphite−
bromine for the noncryogenic preparation of an important vinyl
bromide intermediate 5 in batch mode. We have demonstrated
the use of continuous flow chemistry tools for instantaneous
generation of the brominating reagent, thus avoiding its storage
and potential decomposition prior to reaction with the substrate.
Understanding the instability of the preformed brominating
agent was critical to this process, along with the ability to form
the active agent in situ. We believe that the modified addition
sequence described herein will be generally applicable for the
large scale conversion of a variety of ketones to the corre-
sponding vinyl bromides.
EXPERIMENTAL SECTION
■
First-Generation Batch Process (Figure 1). Molecular
bromine (6.2 L, 121.0 mol) was charged to a solution of
triphenyl phosphite (37.2 kg, 119.9 mol) in dichloromethane
(347 L) in a Hastelloy reactor while maintaining the mass
temperature between −65 and −75 °C. The reaction mixture
was stirred for 30 min at −65 to −75 °C, and triethylamine
Table 1. Comparison of First- and Second-Generation Batch Processes
first-gen batch process
second-gen batch process
(Figure 4)
advantages of second-gen batch
process
parameter
(Figure 1)
operating temperature
−75 to 30 °C
one
−20 to 30 °C
none
cryogenic conditions are not
required
number of cryogenic reactors required (operating below −50 °C)
reaction time (after addition of reagents)
18−24 h
∼1 h
significant reduction in
reaction time
cycle time/batch (500 g scale)
yields
∼3 days
55−74%
1 day
∼66% reduction in cycle time
up to 35% yield improvement
85−90%
D
DOI: 10.1021/acs.oprd.6b00100
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Analytical Data of Purified 5. IR (cm−1): 2933, 1727,
1652, 1168, 1030.
(9) The reaction was extremely rapid at this temperature. We have
not investigated the exact nature of the brominating species generated
under these conditions as part of this work.
1H NMR (400 MHz, DMSO-d6): δ 6.04 (s, 1 H), 4.08
(q, J = 7.2 Hz, 2 H), 2.58−2.61 (m, 1 H), 2.38−2.51 (m, 2 H),
2.19−2.33 (m, 2 H), 1.96−2.0 (m, 1 H), 1.72−1.81 (m, 1 H),
1.19 (t, J = 7.0 Hz, 3 H).
13C NMR (100 MHz, DMSO-d6): δ 173.81, 127.17, 120.95,
59.93, 36.81, 33.53, 29.00, 26.31, 13.94.
AUTHOR INFORMATION
Corresponding Author
*E-mail: vaidy@bms.com.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Jayaprakash Karamil, Sabuj Mukherjee, and Tamas
Benkovics for providing chemistry insights, Lakshmi Narasimhan
and Somprabha Sidar for process safety analysis, and Murali
Arumugam and Santosh Gandhi for analytical support. We
would also like to acknowledge David Kronenthal, Ke Chen,
and Robert Waltermire for their support of this work.
REFERENCES
■
(1) (a) Biajoli, A. F. P.; Schwalm, C. S.; Limberger, J.; Claudino, T.
S.; Monteiro, A. L. J. Braz. Chem. Soc. 2014, 25, 2186. (b) Magano, J.;
Dunetz, J. R. Transition Metal-Catalyzed Couplings in Process Chemistry:
Case Studies From the Pharmaceutical Industry; Wiley-VCH, 2013.
(c) Johansson Seechurn, C. C. C.; Kitching, M. O.; Colacot, T. J.;
Snieckus, V. Angew. Chem., Int. Ed. 2012, 51, 5062. (d) Nicolaou, K.
C.; Bulger, P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44, 4442.
(e) Espinet, P.; Echavarren, A. M. Angew. Chem., Int. Ed. 2004, 43,
4704. (f) Kotha, S.; Lahiri, K.; Kashinath, D. Tetrahedron 2002, 58,
9633.
(2) (a) Kamei, K.; Maeda, N.; Ogino, R.; Koyama, M.; Nakajima, N.;
Tatsuoka, T.; Ohno, T.; Inoue, T. Bioorg. Med. Chem. Lett. 2001, 11,
595. (b) Eszenyi, T.; Tímar, T.; Sebok, P. Tetrahedron Lett. 1991, 32,
́
́
̃
827. (c) Fry, A. J.; Moore, R. H. J. Org. Chem. 1968, 33, 425. (d) Hurd,
C. D.; Hayao, S. J. Am. Chem. Soc. 1954, 76, 5065.
(3) (a) Kamei, K.; Maeda, N.; Tatsuoka, T. Tetrahedron Lett. 2005,
46, 229. (b) Barton, D. H. R.; O’Brien, R. E.; Sternhell, S. J. Chem. Soc.
1962, 470. (c) Pross, A.; Sternhell, S. Aust. J. Chem. 1970, 23, 989.
(d) Campbell, J. R.; Pross, A.; Sternhell, S. Aust. J. Chem. 1971, 24,
1425. (e) Barton, D. H. R.; Bashiardes, G.; Fourrey, J.-L. Tetrahedron
1988, 44, 147. (f) Barton, D. H. R.; Bashiardes, G.; Fourrey, J.-L.
Tetrahedron Lett. 1983, 24, 1605. (g) Furrow, M. E.; Myers, A. G. J.
Am. Chem. Soc. 2004, 126, 5436.
(4) Spaggiari, A.; Vaccari, D.; Davoli, P.; Torre, G.; Prati, F. J. Org.
Chem. 2007, 72, 2216.
(5) Quenching the reaction mixture at − 60 °C led to only 5% of
isolated product. Quenching of the reaction immediately after
warming to room temperature led to ca. 40% yield of 5.
(6) The potency of the product in the crude was determined by
HPLC assay (typical range: 28−30%), and the yield of the reaction
was corrected for the potency of product 5 as well as starting material
3. The subsequent base mediated hydrolysis of ester 5 to the
corresponding carboxylic acid led to complete cleanup of the product
upon acid−base workup and purification of the crude by water slurry.
(7) (a) Hatfield, L. D.; Blaszczak, L. C.; Fisher, J. W. U.S. Patent
4,230,644, 1980. (b) Bingham, A., Jr. U.S. Patent 4,240,988, 1980.
(c) Michalski, J.; Mikolajczak, J.; Skowronska, A. J. Am. Chem. Soc.
1978, 100, 5386. (d) Michalski, J.; Pakulski, M.; Skowronska, A. J. Org.
Chem. 1980, 45, 3122. (e) Spaggiari, A.; Blaszczak, L. C.; Prati, F. Org.
Lett. 2004, 6, 3885.
(8) These flow rates correspond to a 1:1.1 mol ratio of bromine to
triphenylphosphite.
E
DOI: 10.1021/acs.oprd.6b00100
Org. Process Res. Dev. XXXX, XXX, XXX−XXX