Safe and Efficient Decarboxylation Process: A Practical Synthetic Route to 4-Chlorobenzo[b]thiophene

Article abstract of DOI:10.1021/acs.oprd.5b00340

We established an improved synthetic route to 4-chlorobenzo[b]thiophene, a key intermediate in brexpiprazole synthesis, via a practical decarboxylation process in three steps. Thermal analysis demonstrated that the coexistence of the decarboxylated product with DBU should be avoided and that removal of the product outside the reactor was vital. Our process yields the target compound by distillation under reduced pressure and is safe, highly batch efficient, cost-effective, and high yielding. Furthermore, manufacturing on a pilot scale was also accomplished through our approach.

Full text of DOI:10.1021/acs.oprd.5b00340

Communication
pubs.acs.org/OPRD
Safe and Efficient Decarboxylation Process: A Practical Synthetic
Route to 4‑Chlorobenzo[b]thiophene
Masahiro Miyake,* Masamichi Shimizu, Koichi Tsuji, and Keiichi Ikeda
Bulk Pharmaceutical Chemicals Department, Second Tokushima Factory, Production Headquarters, Otsuka Pharmaceutical Co., Ltd.,
224-18, Hiraishi Ebisuno, Kawauchi-cho, Tokushima 771-0182, Japan
S
* Supporting Information
are unfavorable due to residual amounts in active pharmaceut-
ABSTRACT: We established an improved synthetic route
to 4-chlorobenzo[b]thiophene, a key intermediate in
brexpiprazole synthesis, via a practical decarboxylation
process in three steps. Thermal analysis demonstrated that
the coexistence of the decarboxylated product with DBU
should be avoided and that removal of the product outside
the reactor was vital. Our process yields the target
compound by distillation under reduced pressure and is
safe, highly batch efficient, cost-effective, and high yielding.
Furthermore, manufacturing on a pilot scale was also
accomplished through our approach.
ical ingredients (APIs). In particular, residual quinoline should
be strictly controlled because of genotoxic issues, and therefore,
alternative strategies are recommended. While there have been
several publications reporting decarboxylation by other
methods, including the use of copper/phenanthroline,6 gold,7
and silver,6a,8 we focused on decarboxylation using DBU.2c,9
While this method is advantageous in not requiring metals or
genotoxic chemicals, it does have some drawbacks for its
application to large-scale production. For example, this method
sometimes requires special apparatus, such as microwaves or
tube reactors, to boost reaction rates. In addition, solvents with
higher boiling points must be used since the reaction normally
occurs at higher temperature. Especially, these solvents are
difficult to remove from the reaction mixture, leading to the
problem of residual solvents in APIs. Thus, these issues result
in high costs, batch inefficiency, and a large heat capacity. To
resolve these issues, we believed that carrying out this reaction
in the absence of solvents would be ideal. We therefore report
the successful synthesis of 4-chlorobenzo[b]thiophene 2 from
2,6-dichlorobenzaldehyde 3 with a practical decarboxylation via
intermediate 5 (Scheme 1, Our Method).
INTRODUCTION
■
Brexpiprazole 1 is a serotonin−dopamine activity modulator
(SDAM) for the treatment of schizophrenia and is used as
adjunctive therapy for the treatment of clinical depression.1 Its
new drug application (NDA) has recently been approved by
the US Food and Drug Administration (FDA), and clinical
trials for its application in attention deficit hyperactivity
disorder (ADHD), post-traumatic stress disorder (PTSD),
and agitation associated with dementia of the Alzheimer’s type
are currently underway.
RESULTS AND DISCUSSION
■
We first examined the chemical transformation of 3 to 4. Based
on a previously reported method,2b we attempted the reaction
of 3 with 1.3 eq of TGA and 3 eq of aqueous KOH under
reflux. After 3 h, the resulting mixture gave the target
compound 4 in 62% isolated yield. A similar result was
obtained when KOH was replaced with NaOH. Since the air
oxidation of TGA and sublimation of 3 during the reaction
hindered an improvement in yield, we carried out the reaction
under N2 using a mixture of DME/water as solvent to dissolve
sublimed 3. Following reaction optimization, 4 was obtained in
the highest yield (82%) with the use of 1.3 eq of TGA and 3 eq
of aqueous NaOH in water/DME (7:3) under reflux.
In the preparation of brexpiprazole 1, 4-chlorobenzo[b]-
thiophene 2 is a key intermediate, and therefore, cost reduction
for the synthesis of 2 is directly related to the cost of
brexpiprazole production.2 To achieve a more efficient
synthetic route to 2, the selection of an appropriate starting
material is particularly important. Several approaches to
benzo[b]thiophene analogs have been reported, with reactions
starting from benzaldehydes and thioglycolic acid (TGA) being
regarded as practical and economical because of the wide
availability of both materials.2b,3 Of these approaches, we
considered the preparation of 2 from 2,6-dichlorobenzaldehyde
3, which was considered a desirable starting material in terms of
its regioselectivity and low material expenses (Scheme 1,
Conventional Method).4 This route requires two trans-
formations, namely, the formation of benzothiophene carbox-
ylic acid 4 and its subsequent decarboxylation. Given the lower
material cost of 3 and fewer number of steps to 2, this strategy
was deemed reasonable.
We subsequently investigated the decarboxylation of 4 for
the preparation of 2, beginning with the neat reaction of 4 with
DBU under Ar atmosphere at 200 °C. Although the reaction
reached completion using 1 eq or even a catalytic amount of
DBU, the resulting mixture turned black, and decolorization of
2 was problematic. Since we assumed that oxygen was
responsible for causing this color, these reactions were repeated
under Ar to exclude oxygen. However, no improvement was
Generally, in the synthesis of benzothiophene analogues
from their corresponding carboxylic acids, decarboxylation
using copper/quinoline is widely utilized, due to its ease of
handling and the availability of each material.5 However, this
general procedure demands copper metal and quinoline which
Received: October 25, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.5b00340
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Scheme 1. Proposed Synthetic Route to Thiophene 2 from Aldehyde 3
Figure 1. DSC analysis of the relevant compounds.
observed. This result was disappointing, as for APIs, color
should generally be avoided.
methanol, its high solubility to methanol resulted in a drop in
yield following crystallization. It was therefore necessary to
evaporate methanol at high temperature in combination with
solvents of higher boiling point. Butyl acetate was found to be
the optimal solvent for this procedure, and methanol was
successfully removed to give 5 in high yield (∼98%).
To understand the characteristics of the decarboxylation
reaction, we conducted differential scanning calorimetry (DSC)
measurements on 2 mg of sample under Ar. An aluminum pan
with a pierced lid was used, along with a linear heating rate of 3
°C/min. When 4 and DBU were analyzed individually, no
conspicuous peaks regarding exotherm were observed (Figure
1, left graph). To evaluate the reaction system, a 1:1 molar ratio
mixture of 4 and DBU (5, DBU salt of 4) was then prepared.
Although two endothermic peaks representing the melting
point of 5 and the reaction heat were detected, an exothermic
peak was also observed at 320 °C (Figure 1, left graph). From
the exothermic onset temperature obtained here, the ADT24
(the temperature at which TMRad is 24 h, as derived from
adiabatic measurements) of 5 was calculated to be 162 °C,
In the subsequent decarboxylation, compound 5 was placed
in the reactor and heated to 190−200 °C under reduced
pressure. When the external temperature reached 180 °C, 5
began to melt, changing from a solid to a viscous liquid.
Distillation of this viscous liquid gave a colorless distillate,
which contained compound 2. As the distillate was a mixture of
2 with DBU, it was acidified using HCl to remove DBU.11
Following extraction with ethyl acetate, the solution was
concentrated to afford 2 as a pure colorless oil in high yield.
Our strategy therefore appeared to be the only successful
approach for preventing the coexistence of decarboxylated 2
and DBU within the reaction system, by expelling the mixture
by distillation. In addition, our method is superior as it allows
simultaneous purification and decarboxylation. Our results also
confirmed that the coloring was not caused by oxygen but by an
intermolecular reaction between 2 and DBU.12
To acquire safety information for our process, we
investigated the relationship between temperature and pressure.
When the reaction was carried out at 7.2 kPa, the vapor
temperature of 2 during the distillation was 152 °C, which was
less than the calculated ADT24 value. Actually, 7.2 kPa is
feasible pressure also on large scale; therefore, it was found that
this decarboxylation could be performed safely. As a result, no
colored material was observed in the reactor or in the distillate,
and we established a safer pyrogenic reaction that allowed
simultaneous distillation and decarboxylation. Furthermore, this
based on the report by Hungerbuhler et al.10 This indicated
̈
that the decarboxylation reaction at 200 °C could not be
considered safe.
To investigate this in further detail, additional measurements
were carried out. Surprisingly, although the DSC result from 2
did not show a noticeable exothermic peak, a similar
exothermic peak to 5 was observed in the mixture of 2 with
DBU (Figure 1, right graph). This suggests that decarboxylated
product 2 reacted with DBU to generate heat, meaning that the
two should not be present in the same reactor at high
temperature. As we found that the DBU salt 5 could be isolated
as a crystal, we herein proposed a method where 4 was
transformed into the DBU salt 5, followed by heating of
intermediate 5 under reduced pressure in the absence of
solvents, to give 2 following distillation.
We then focused on optimization for the synthesis of
compound 5. While the formation of 5 proceeded smoothly in
B
DOI: 10.1021/acs.oprd.5b00340
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Scheme 2. Established Synthetic Route to 2
was also a highly batch efficient process, which did not require
additional reagents or solvents.
and the mixture was degassed over three vacuum/nitrogen
cycles. TGA (13.7 kg, 147.89 mol) was then added, and the
resulting mixture stirred under a flow of N2. After 15 min, 2,6-
dichlorobenzaldehyde 3 (19.91 kg, 113.76 mol) was added and
heated at reflux for 9 h (N2 flow continued until just before
reflux). The mixture was then cooled and stirred at <5 °C for 1
h. The resulting precipitate was filtered and washed with
purified water (60 L) to give a white solid, which was dissolved
in purified water (200 L) and heated to 80 °C. Activated carbon
(0.4 kg) was then added and the mixture stirred at 90 °C for 30
min. The resulting suspension was filtered at 80−90 °C,
following the addition of purified water (20 L). Following
filtration, the filtrate was heated to >70 °C, conc. HCl (11.7 kg,
113.76 mol) and purified water (11.7 L) were added, and
stirring was carried out at 75 °C for 15 min. The slurry was
then cooled <30 °C, filtered, and washed with purified water
(200 L). The obtained wet cake was dried at 80 °C to give the
desired product 4 as a white solid (19.85 kg, 82.06%). mp 260
With the optimized reaction conditions in hand, we
manufactured 2 on a pilot scale starting from ∼20 kg of 3
(Scheme 2). Compound 3 was heated at reflux in the presence
of TGA and 25% aqueous NaOH under N2 in a mixed solvent
(water/DME) to afford 4 as its sodium salt. Treatment with
HCl afforded 4 in good yield (82%). The reaction of 4 with
DBU was then achieved, followed by concentration of the
reaction mixture using butyl acetate, allowing the elimination of
methanol. Cooling of the resulting slurry afforded 5 in excellent
yield (∼98%). Finally, 5 was heated under reduced pressure,
and the distillate was washed and concentrated to give pure 2 as
a colorless oil in excellent yield (18.69 kg, ∼ 93%). This gave an
overall yield of approximately 75% over three steps.
CONCLUSION
■
A practical route to the preparation of 4-chlorobenzo[b]-
thiophene 2, a key intermediate in the synthesis of
brexpiprazole, was developed. Our synthetic route began from
2,6-dichlorobenzaldehyde, a low cost material, to give target
compound 2 in good yield over three steps. Thermal analysis
by DSC offered crucial information regarding the decarbox-
ylation step to allow the reaction to proceed safely. In addition,
our decarboxylation process allowed the distillation of the
decarboxylated compound under reduced pressure to give the
colorless and pure 2. Our method fulfills a number of key
process requirements, including safety, high batch efficiency,
high yield, and low cost. Finally, using this method, we
succeeded in manufacturing 2 on a pilot scale. Following
further optimization, this procedure should be applicable on a
commercial scale. Studies into the scale up of this process for
commercial production are currently underway.
1
°C. H NMR (300 MHz, DMSO-d6) δ 7.54 (d, 1H, J = 11.6,
7.7 Hz), 7.56 (dd, 1H, J = 17.8, 7.7 Hz), 8.03 (d, 1H, J = 0.7
Hz), 8.07 (td, 1H, J = 7.6, 0.9 Hz), 13.19 (brs, 1H). 13C NMR
(75 MHz, DMSO-d6) δ 122.21, 125.01, 126.84, 127.99, 128.96,
136.50, 136.57, 142.55, 163.10. Elemental analysis calcd for C:
50.83%, H: 2.37%, found C: 50.84%, H: 2.21%.
2,3,4,6,7,8,9,10-Octahydropyrimido[1,2-a]azepin-1-
ium 4-Chlorobenzo[b]thiophene-2-carboxylate 5. A
mixture of 4 (19.81 kg, 93.16 mol), methanol (40 L), and
DBU (14.18 kg, 93.16 mol) was stirred to give a homogeneous
solution. Butyl acetate (99 L) was then added, and the mixture
was heated to 110 °C to remove methanol by evaporation. The
mixture was cooled to <25 °C, butyl acetate was added (19.8
L), and stirring continued at <25 °C for 3 h. The resulting
slurry was filtered and washed with butyl acetate (29.6 L) to
give a wet cake, which was dried at 80 °C to give 5 as a white
1
solid (33.31 kg, 97.97%). Mp 182.5 °C. H NMR (300 MHz,
EXPERIMENTAL SECTION
CDCl3) δ 1.66 (m, 6H), 1.80−1.75 (m, 2H), 2.98−2.94 (m,
2H), 3.45−3.39 (m, 4H), 3.55−3.51 (m, 2H), 7.31−7.20 (m,
2H), 7.69 (dd, 1H, J = 3.9, 0.6 Hz), 7.97 (s, 1H), 13.19 (brs,
1H). 13C NMR (75 MHz, CDCl3) δ 19.51, 24.03, 26.70, 28.88,
31.99, 38.01, 48.36, 53.94, 121.04, 123.00, 125.17, 126.61,
128.98, 138.21, 142.43, 146.85, 165.91, 167.20. Elemental
analysis calcd for C: 59.25%, H: 5.80%, N: 7.68%, found C:
59.10%, H: 5.44%, N: 7.53%.
■
All commercially available materials and solvents were used as
received without any further purification. The conversion of 3
to 4 was monitored by high performance liquid chromatog-
raphy (HPLC) analysis using a Shimadzu SLC 10Avp System.
NMR spectra were recorded on a Bruker-Spectrospin 300 with
TMS (tetramethylsilane) as an internal standard in DMSO-d6/
CDCl3. Chemical shifts are reported in parts per million (ppm)
on the δ scale from TMS. The purity of 2 was measured by gas
chromatography using an Agilent technology GC 6890 System.
Differential scanning calorimetry (DSC) was performed using a
Shimadzu DSC-60A system.
4-Chlorobenzo[b]thiophene (2). 5 (43.58 kg, 119.43
mol) was placed in the reactor, heated to 190−200 °C, and
distilled under reduced pressure. Ethyl acetate (131 L) and
purified water (44 L) were added to the obtained distillate. The
aqueous layer was removed, and a further portion of purified
water (22 L) and conc. HCl (1.31 kg, 0.03 wt %) were added to
the organic layer. After washed with purified water (33 L), the
4-Chlorobenzo[b]thiophene-2-carboxylic Acid (4). Pu-
rified water (95.9 L), 1,2-dimethoxyethane (20.9 kg), and 25%
NaOH aq. (54.9 kg, 341.29 mol) were added to the reactor,
C
DOI: 10.1021/acs.oprd.5b00340
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
(4) $95/kg for 2,6-dichlorobenzaldehyde and $33/kg for TGA
(5) Recent publications concerning decarboxylation of benzothio-
phene analogues using Cu/quinoline, see: (a) Wang, G.; Beigelman,
L.; Truong, A.; Napolitano, C.; Andreotti, D.; He, H. Compounds for
the Treatment of Paramoxyvirus Viral Infections.
WO2014031784(A1), Feb 27, 2014. (b) Li, C.-Y.; Su, C.; Wang, H.-
H.; Kumaresan, P.; Hsu, C.-H.; Lee, I.-T.; Chang, W.-C.; Tingare, Y.
S.; Li, T.-Y.; Lin, C.-F.; Li, W.-R. Dyes Pigm. 2014, 100, 57−65.
(c) Niculescu-Duvaz, D.; Niculescu-Duvaz, I.; Suijkerbuijk, B. M. J. M.;
residual ethyl acetate solution was evaporated at atmospheric
pressure then reduced pressure to give a 2 as a colorless oil
1
(18.69 kg, 92.78%) in >99.9% purity. H NMR (300 MHz,
CDCl3) δ 7.26 (t, 1H, J = 7.8 Hz), 7.36 (dd, 1H, J = 7.8, 0.9
Hz), 7.50 (d, 1H, J = 5.7 Hz), 7.52 (d, 1H, J = 5.7 Hz), 7.76 (d,
1H, J = 7.8 Hz). 13C NMR (75 MHz, CDCl3) δ 121.12, 122.40,
125.02, 127.43, 128.93, 138.06, 141.07. Elemental analysis calcd
for C: 56.98%, H: 2.99%, found C: 56.76%, H: 2.94%.
Men
R.; Springer, C. J. Bioorg. Med. Chem. 2013, 21, 1284−1304.
(6) (a) Rudzki, M.; Alcalde-Aragones, A.; Dzik, W. I.; Rodríguez, N.;
Gooßen, L. J. Synthesis 2012, 2, 184−193. (b) Cadot, S.; Rameau, N.;
Mangematin, S.; Pinel, C.; Djakovitch, L. Green Chem. 2014, 16,
3089−3097.
(7) Dupuy, S.; Nolan, S. P. Chem. - Eur. J. 2013, 19, 14034−14038.
(8) (a) Lu, P.; Sanchez, C.; Cornella, J.; Larrosa, I. Org. Lett. 2009,
11, 5710−5713. (b) Maehara, A.; Tsurugi, H.; Satoh, T.; Miura, M.
Org. Lett. 2008, 10, 1159−1162. (c) Liu, K.; Jia, F.; Xi, H.; Li, Y.;
Zheng, X.; Guo, Q.; Shen, B.; Li, Z. Org. Lett. 2013, 15, 2026−2029.
(d) Toy, X. Y.; Roslan, I. I. B.; Chuah, G. K.; Jaenicke, S. Catal. Sci.
Technol. 2014, 4, 516−523.
(9) For selected papers, see: (a) Owton, W. M. Tetrahedron Lett.
2003, 44, 7147−7149. (b) Tilstam, U. Org. Process Res. Dev. 2012, 16,
1449−1454.
́
ard, D.; Zambon, A.; Davies, L.; Pons, J. F.; Whittaker, S.; Marais,
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.oprd.5b00340.
■
́
S
Characterization spectra (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: m_miyake@otsuka.jp.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(10) Pastre,
Prev. Process Ind. 2000, 13, 7−17.
(11) Boiling point for DBU; 97−98/3 Torr see FR1542058, Oct 11,
1968.
́
J.; Worsdorfer, U.; Keller, A.; Hungerbuhler, K. J. Loss
̈ ̈
̈
We thank Ms. Shinohara for elemental analysis and K.
Watanabe, H. Matsumoto, J. Fukui, M. Yoshida, and H.
Morimoto for their contribution to this research.
(12) One-electron reduction by single electron transfer of DBU
might happen. Indeed, the decarboxylation of 4-chloro-2-nitrobenzoic
acid with DBU under argon atmosphere furnished 3-chloroaniline
dominantly. In contrast, no aniline compound was observed in the
same reaction under reduced pressure. Single electron transfer of DBU
had been previously reported, see: Skiebe, A.; Hirsch, A.; Klos, H.;
Gotschy, B. Chem. Phys. Lett. 1994, 220, 138−140.
ABBREVIATIONS
■
SDAM, serotonin-dopamine activity modulator; NDA, new
drug application; FDA, Food and Drug Administration;
ADHD, attention-deficit hyperactivity disorder; PTSD, post-
traumatic stress disorder; TGA, thioglycolic acid; DME,
dimethoxyethane; API, Active Pharmaceutical Ingredients;
DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; DSC, differential
scanning calorimetry; ADT24, temperature at which TMRad is
24 h derived from adiabatic measurements; TMRad, time to
maximum rate under adiabatic conditions
REFERENCES
■
(1) (a) Correll, C. U.; Skuban, A.; Ouyang, J.; Hobart, M.; Pfister, S.;
McQuade, R. D.; Nyilas, M.; Carson, W. H.; Sanchez, R.; Eriksson, H.
Am. J. Psychiatry 2015, 172, 870−880. (b) Maeda, K.; Lerdrup, L.;
Sugino, H.; Akazawa, H.; Amada, N.; McQuade, R. D.; Stensbøl, T. B.;
Bundgaard, C.; Arnt, J.; Kikuchi, T. J. Pharmacol. Exp. Ther. 2014, 350,
605−614.
(2) Synthetic routes to 4-chlorobenzo[b]thiophene, see: (a) Hansen,
M. M.; He, J.-X.; Honigschmidt, N. A.; Koch, D. J.; Kohn, T. J.; Rocco,
V. P.; Spinazze, P. G.; Takeuchi, K. Indole Derivatives for the
Treatment of Depression and Anxiety. US2004006229(A1), Jan 8,
2014. (b) Shinhama, K.; Utsumi, N.; Sota, M.; Fujieda, S.; Ogasawara,
S. Method for Producing Benzo[b]thiophene Compound.
WO2013015456, Jan 31, 2013. (c) Allen, D.; Callaghan, O.;
Cordier, F. L.; Dobson, D. R.; Harris, J. R.; Hotten, T. M.; Owton,
W. M.; Rathmell, R. E.; Wood, V. A. Tetrahedron Lett. 2004, 45, 9645−
9647. (d) Ewing, W. R. et al. Substituted Oxoazaheterocyclyl
Compounds. US2004102450(A1), May 27, 2004. (e) Doyle K. J.;
Kerrigan F.; Watts J. P. Dihydroimidazo[2,1-b]thiazole and Dihydro-
5H-thiazolo[3,2-a]pyrimidines as Antidepressant agents.
US2003166628(A1), Sep 4, 2003. (f) Hansch, C.; Schmidhalter, B.
J. Org. Chem. 1955, 20, 1056−1061.
(3) (a) Rahman, L. K. A.; Scrowston, R. M. J. Chem. Soc., Perkin
Trans. 1 1984, 3, 385−390. Jiang et al. had recently reported on the
synthesis of 4-(1-piperazinyl)benzo[b]thiophene which was the
component of brexpiprazole, see: (b) Wu, C.; Chen, W.; Jiang, D.;
Jiang, X.; Shen, J. Org. Process Res. Dev. 2015, 19, 555−558.
D
DOI: 10.1021/acs.oprd.5b00340
Org. Process Res. Dev. XXXX, XXX, XXX−XXX