Development of a manufacturing process for sibenadet hydrochloride, the active ingredient of viozan

Article abstract of DOI:10.1021/op049953y

A process for commercial manufacture of the dual D₂- β₂ receptor agonist sibenadet hydrochloride has been developed. The process relies upon introduction of operationally simple chemistry at the final stages where two key intermediates are reacted to assemble the final molecule, isolated by crystallization. A nine-stage sequence for synthesis of the key amine hydrochloride intermediate was developed, and modifications to the original process are described. Major strategic improvements were made in definition of the final route to the "side chain" precursor molecule, the second key intermediate, hinging around a thiyl radical addition and subsequent high-yielding telescoped processes for synthesis of this highly crystalline benzoate ester. Development of these chemistries is discussed, together with some issues surrounding definition of the final validated commercial processes.

Full text of DOI:10.1021/op049953y

Organic Process Research & Development 2004, 8, 628−642
Development of a Manufacturing Process for Sibenadet Hydrochloride, the
Active Ingredient of Viozan
Melvyn E. Giles,* Colin Thomson, Stephen C. Eyley, Andrea J. Cole, Christopher J. Goodwin, Paul A. Hurved,
Andrew J. G. Morlin, James Tornos, Ste´phanie Atkinson, Cecelia Just,† Janet C. Dean, John T. Singleton,
Andrew J. Longton, Ian Woodland, Andrew Teasdale, Bjo¨rn Gregertsen,† Howard Else, Mandip S. Athwal,
Stephen Tatterton, Jason M. Knott, Nicola Thompson, and Stephen J. Smith
AstraZeneca R & D Charnwood, Process R & D, Bakewell Road, Loughborough,
Leicestershire LE11 5RH, United Kingdom
Abstract:
A process for commercial manufacture of the dual D2-â2
receptor agonist sibenadet hydrochloride has been developed.
The process relies upon introduction of operationally simple
chemistry at the final stages where two key intermediates are
Figure 1. Sibenadet Hydrochloride.
reacted to assemble the final molecule, isolated by crystalliza-
tion. A nine-stage sequence for synthesis of the key amine
hydrochloride intermediate was developed, and modifications
to the original process are described. Major strategic improve-
ments were made in definition of the final route to the “side
chain” precursor molecule, the second key intermediate, hinging
Figure 2. Key benzothiazolone amine hydrochloride interme-
around a thiyl radical addition and subsequent high-yielding
telescoped processes for synthesis of this highly crystalline
benzoate ester. Development of these chemistries is discussed,
together with some issues surrounding definition of the final
validated commercial processes.
diate.
Results and Discussion
Route Strategy. All routes investigated employed the
benzothiazolone amine hydrochloride 2 as a key intermediate
(Figure 2). This compound has been described previously
in the literature as a D2 receptor agonist,3 and its modification
by appending a variety of groups to the amino function was
an important element of the medicinal chemistry programme
that led to the discovery of 1. The synthetic strategy that we
embarked upon involved adopting key intermediate 2 as
antecedent to 1 and concentrating efforts on finding an
efficient method for the downstream chemistry. Notwith-
standing this strategy, considerable improvements in the
synthetic route and processing steps to 2 were needed to
ensure adequate throughput.
Route Development for Benzothiazolone Amine Hy-
drochloride 2. The final route used for large-scale prepara-
tion of 2 is outlined in Scheme 1. Initial preparations
employed the known route to trifluoroacetamide 3,3 depro-
tection of which was achieved in three separate steps. Early
on we were able to demonstrate that this circuitous depro-
tection, involving displacement, hydrolysis, dealkylation, and
deacylation, could be streamlined simply by heating 3 with
48% aqueous hydrobromic acid, which effected a similar set
of transformations in a single operation to give the crystalline
hydrobromide salt 2a. Once it had been established that
analogous chemistry was feasible using acetyl protection
instead of trifluoroacetyl for the primary amine function,
efforts were focused on process development of this chem-
istry, culminating in the nine-stage sequence.
Introduction
Sibenadet hydrochloride (1) (Figure 1) is a highly potent
drug designed for treatment of patients with chronic obstruc-
tive pulmonary disease,1,2 a condition for which there is a
definite clinical need, and exerts its biological effects through
a combination of â2 adrenoceptor and D2 dopamine receptor
agonist activities. Whilst the high potency of this compound
via an inhalation route of administration meant that cost of
goods constraints could be relaxed, the predicted size of the
market led us to forecast manufacturing volumes of several
tonnes per year.
Thus, throughput became the main focus of our efforts
in developing a synthetic route suitable for scale-up to
commercial manufacture. The success of these efforts relied
heavily on the introduction of a facile entry to the final
compound, elimination of chromatographic purification, and
definition of efficient chemistry for construction of the side
chain precursor.
* To whom correspondence should be addressed. Fax: +44 0 1509 64 5588.
E-mail: melvyn.giles@astrazeneca.com.
† AstraZeneca R & D So¨derta¨lje, Process R & D, S-15185 So¨derta¨lje, Sweden.
(1) Hiller, F. C.; Alderfer, V.; Goldman, M. Respiratory Medicine 2003, 97
Suppl. A, 45.
(2) Bonnert, R. V.; Brown, R. C.; Chapman, D.; Cheshire, D. R.; Dixon, J.;
Ince, F.; Kinchin, E. C.; Lyons, A. J.; Davis, A. M.; Hallam, C.; Harper, S.
T.; Unitt, J. F.; Dougall, I. G.; Jackson, D. M.; McKechnie, K.; Young, A.;
Simpson, W. T. J. Med. Chem. 1998, 41, 4915.
(3) Weinstock, J.; Gaitanopoulos, D. E.; Stringer, O. D.; Franz, R. G.; Hieble,
P.; Kinter, L. B.; Mann, W. A.; Flaim, K. E.; Gessner, G. J. Med. Chem.
1987, 30, 1166.
628
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development
10.1021/op049953y CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/16/2004

Scheme 1. Synthetic route and conditions used for large-scale production of 2
Of the three stages used to convert 2-methoxyphenyl-
ethylamine to the crystalline aniline hydrochloride salt 4, the
protection and reduction steps were relatively straightforward.
For the nitration stage, 70% nitric acid in acetic acid was
employed initially. Hazard testing concentrated mainly on
the thermal characteristics of the reaction. A Dewar flask
experiment in which the reagents were combined in one
portion showed a 12 °C temperature rise over 15 min, and
when the mixture had reached 35-40 °C, the temperature
increased dramatically reaching a maximum of 103 °C over
the next 10 min with heavy gas evolution. Reaction calo-
rimetry using 10 mole equiv of 70% nitric acid and 3 weight
equiv of acetic acid with addition of substrate to the nitrating
media predicted an adiabatic temperature rise of 36 °C with
accumulation (heat was evolved for an hour after the addition
was complete). The calorimetry showed that reaction begins
as soon as addition of substrate solution to the nitric acid is
started but can stall if the temperature drops below 15 °C.
This meant that accumulation could be minimized by running
the reaction in semibatch mode, with controlled addition of
the substrate dissolved in acetic acid to the nitric acid
between 15 and 25 °C. Replacing nitric acid with acetyl
nitrate prepared in situ as nitrating agent, a modification
introduced by an outsourcing partner, offered considerable
advantages for this stage. This procedure used fewer
equivalents of nitrating agent and was also shown to be safe
to run in the alternative addition mode with fuming nitric
acid being added in small aliquots keeping the reaction
mixture between 15 and 25 °C after each aliquot addition.
These modifications resulted in significant reduction in
byproducts derived from over-nitration and hence enhanced
yield.
Upon attempted thiocarbamoylation of 4, it soon became
clear that formation of the thiourea 6 using simple thiocy-
anate salts was problematic, being seriously compromised
by formation of significant amounts of the symmetrical
thiourea derived from two molecules of 4. During scale-up,
we were only able to avoid this by introducing a protecting
group, reacting the free base obtained from 4 with benzoyl
isothiocyanate prepared in situ from ammonium thiocyanate
and benzoyl chloride. This gave the benzoyl thiourea 5,
isolable by precipitation and filtration, and a deprotection
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
629

Table 1. Conditions investigated for oxidative cyclization of 6 using 1 equiv of NBS
isolated
yield of 7
(%)
HPLC
purity
(%)
scale
(mmol of 6)
solvents
conditions
15
37
37
94
1.9
TFA
TFA
MeSO3H
MeSO3H + 10% v/v AcOH
AcOH + 2 equiv of H2SO4
ambient (NBS addition)-60 °C 30 min
ambient (NBS addition)-60 °C 30 min
5 °C (NBS addition)-ambient 5 hr
5 °C (NBS addition)-ambient 5 hr
ambient temperature 21 h
87
82
81
96
94
92
91
94
88
52 (oil precipitated)
step was now required to liberate the cyclization precursor
6. Fortunately this was easily achieved by treatment with
alkali, and the two-step sequence afforded material of good
quality and in 81% isolated overall yield.
intermediate 8 defined the development work needed for the
substitution of amino for chloro via diazotization. Optimal
conditions for this process involved controlled addition of
aqueous sodium nitrite to a solution of 7 in concentrated
hydrochloric acid containing defined sub-stoichiometric
quantities of copper(I) and copper(II) chloride. It was found
that controlling the temperature during dissolution and
reaction of 7 as well as the reaction time minimized
formation of an impurity identified by HPLC-MS as the
chloroethyl derivative 8a. Whilst it was shown that a solution
of the starting material in the reaction acid was stable at 25
°C overnight, heating at 45 °C for 2 h before recooling and
submitting to the reaction conditions resulted in formation
of 5-8% 8a. However, 8a was also produced on extended
reaction times (>3 h) at normal reaction temperature (<18
°C), suggesting that there may be more than one possible
pathway for its formation (i) by reaction of chloride ion with
the primary diazonium salt formed as a result of partial
hydrolysis of the acetyl group and (ii) through the interme-
diacy of the N-nitroso amide in the reaction medium leading
to the same diazonium salt. Although 8a could be effectively
removed during workup, the temperature during dissolution
of 7 was kept below 30 °C and the sodium nitrite addition
time was limited to 2.5 h in order to minimize this side
reaction.
In addition, we found that evolution of nitrogen from the
process caused a certain amount of frothing in the reaction
vessel. Unless care was taken to introduce the sodium nitrite
solution below the surface of the substrate solution, signifi-
cant decomposition of the nitrite could ensue as contact with
acidic froth occurred, necessitating addition of further
quantities of nitrite to achieve good conversions. In the
laboratory, quenching the reaction mixture into water did
not always give reliable precipitation of the product 8 as a
solid, and oiling was sometimes encountered. On the pilot
plant, this was overcome by using a two phase quench
mixture comprising defined quantities of water and methyl
isobutyl ketone, which gave reliable precipitation and also
improved the purity of 8 somewhat. This stage produced the
largest number of impurities of any of the nine-stage
sequence, with the major one (2-4%) identified by HPLC-
MS and NMR as the bis-benzothiazolone Bis-8, which was
shown to carry through to an analogous bis-benzothiazolone
hydrobromide Bis-2a, removed by filtration during conver-
sion of 2a to the hydrochloride 2 (vide infra). Although we
considered the diazotization process operable on a large scale,
the volume efficiency of the stage was low and could benefit
from further optimization.
Next came the oxidative cyclization of 6 to furnish the
benzothiazole nucleus as the 2-amino benzothiazole deriva-
tive 7. N-Bromosuccinimide served as a suitable replacement
for bromine, allowing good control in the charging stoichi-
ometry. This proved crucial in controlling the formation of
a ring-brominated byproduct, a troublesome impurity that is
not removed downstream. Choice of solvents for the cy-
clization stage was limited by the tendency for precipitate
formation prior to complete conversion during this process,
and acidic media were found to alleviate this problem. The
reaction worked well in neat methanesulfonic acid, but
concerns over bulk availability prompted a search for an
alternative. A number of acidic media combinations were
investigated in the laboratory, and selected data are shown
in Table 1.
Although TFA was suitable in the laboratory, its volatility
and corrosive potential precluded its use at scale, and whilst
the sulfuric/acetic acid system looked promising, the isolated
yield and purity of 7 from these reactions did not rival those
from the methanesulfonic acid system and the reactions were
attended by significant oiling out of material hindering
product isolation. This unsuccessful attempt at methane-
sulfonic acid replacement represented an opportunity for
subsequent process improvement but was not addressed
further during activities described herein. Because of the
tendency for the solvent to freeze in neat methanesulfonic
acid (mp 20 °C) runs at around 5 °C, we investigated
cosolvents, and ultimately a mixture comprising methane-
sulfonic acid and around 10% v/v acetic acid proved best.
The proportion of acetic acid was increased to around 20%
v/v during scale-up trials without significant impact on yield
or quality. A slight deficiency of NBS was employed to avoid
formation of the ring-brominated impurities in the reaction,
and the product was precipitated upon basic quench and
isolated in around 84% yield.
A few unsuccessful attempts were made to convert 7 into
benzothiazolone derivative 2 directly, using known hydrolytic/
diazotization protocols. However, conditions found to modify
the 2-amino group (e.g., sodium nitrite in a mixture of
hydrochloric acid, formic acid, and acetic acid4) also led to
significant decomposition during attempted hydrolysis of the
putative diazonium salt, and recourse to the 2-chloro
(4) Jones, G. H.; Venuti, M. C.; Alvarez, R.; Bruno, J. J.; Berks, A. H.; Prince,
A. J. Med. Chem. 1987, 30, 295.
630
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

Scheme 2. Initial route to 1 via the amide intermediate
derived from 2
The final deprotection stage that converted 8 into 2a was
accomplished by heating in 48% aqueous hydrobromic acid.
Efficient scrubbing of the methyl bromide byproduct (bp 4
°C) was required on pilot plant scale to avoid release into
the atmosphere. A mixture of 13% w/w ammonia (density
0.880) and 87% methanol proved the most suitable medium
for scrubbing the gas. Aqueous ammonia alone and sodium
hydroxide solution did not remove the gas. In the pilot plant,
the gaseous reaction effluent was passed sequentially through
the ammonia solution and then through a dilute hydrochloric
acid scrubber solution (via appropriately placed “trap”
vessels) to prevent ammonia release.
Although smooth deprotection/hydrolysis was observed,
it was important to avoid an extended reaction time;
otherwise appreciable decomposition could occur giving rise
to impurities that were not easily removed. It was found that
the conversion from hydrobromide to hydrochloride salt by
recrystallization from hydrochloric acid in the presence of
charcoal was an effective means of purification, and a second
recrystallization then provided material of acceptable quality.
These processing stages were run successfully at pilot
plant scale providing multikilogram quantities of hydro-
chloride 2.
Route Selection for Synthesis of Sibenadet Hydrochlo-
ride: The Amide Route. With a supply route secured for
benzothiazolone amine hydrochloride 2, the definition of
robust downstream chemistry for conversion of 2 to sibenadet
hydrochloride was required. The early small quantities
prepared during the discovery phase employed a route
utilizing an amide intermediate that was subjected to borane-
mediated reduction, outlined in Scheme 2.2 To trial this
chemistry in the laboratory, a supply of the requisite acid 9
was needed, previously synthesized from the known oxa-
thiane 116 via thiol 122 within the medicinal chemistry
department (Scheme 3).
The experience gained from running this sequence in the
laboratory suggested that it was not ideal as the basis of a
supply route to 9, even on the laboratory scale. Two
disadvantages were the use of the unstable phenylacetalde-
hyde as starting material, which is prone to polymerization,
and the supply of 3-mercaptopropanol, which was not
available in bulk quantities. In addition, the calcium/liquid
ammonia reductive cleavage5 of oxathiane 11 was unreliable,
sometimes giving very low yields of thiol 12.
from toluene yielded the pure acid 9 in around 60% overall
yield, and this route was found to be suitable for the
preparation of 9 in ∼50 g quantities in the laboratory.
Although a number of development issues remained to be
addressed should this route become adopted in the longer
term (e.g., conditions for the oxidation of 13), it represented
a considerable improvement over the initial sequence.
Moreover, the thiyl radical addition tactic not only proved
useful within medicinal chemistry in their ongoing analogue
synthesis programme but would also feature in the final
sequence to sibenadet hydrochloride.
Having prepared sufficient quantities of the acid 9, we
turned our attention to the formation of the amide and its
subsequent reduction. It soon became clear that the perfor-
mance of these reactions was hampered by a number of
drawbacks. First, activation of the acid 9 was required in
order to form the amide upon reaction with the amine;
We were able to eliminate these concerns by implement-
ing an alternative route based on a thiyl radical addition
reaction7 after minimal process research effort. Addition of
the thiyl radical derived from mercaptoacetic acid to the allyl
ether 14,8 readily prepared by phase transfer mediated
alkylation of 2-phenyl ethanol, gave the thioether acid 13
together with regioisomer 13a as a byproduct (∼10 mol %).
Subsequent oxidation and purification by recrystallization
(7) (a) Meltzer, P. C.; Liu, S.; Blanchette, H. S.; Blundell, P.; Madras, B. K.
Bioorg. Med. Chem. 2002, 10, 3583. (b) Buskas, T.; Soederberg, E.;
Konradsson, P.; Fraser-Reid, B. J. Org. Chem. 2000, 65, 958. (c) Curci,
M.; Schneider, M.; Mieloszynski, J. L.; Paquer, D. Sulfur Lett. 1992, 15,
169. (d) Hazato, A.; Tanaka, T.; Watanabe, K.; Bannal, K.; Toru, T.;
Okamura, N.; Manabe, K.; Ohtsu, A.; Kamimoto, F.; Kurozumi, S. Chem.
Pharm. Bull. 1985, 33, 1815. (e) Smith, B.; Hernestam, S. Acta Chem. Scand.
1954, 8, 1111.
(5) (a) Kurth, M. J.; Tahir, S. H.; Olmstead, M. M. J. Org. Chem. 1990, 55,
2286. (b) Newman, B. C.; Eliel, E. L. J. Org. Chem. 1970, 35, 3641.
(6) Fuji, K.; Ueda, M.; Sumi, K.; Kajiwara, K.; Fujita, E.; Iwashita, T.; Miura,
I. J. Org. Chem. 1985, 50, 657. This paper describes a low-yielding
preparation of 11 by alkylation of 2-lithio-1,3-oxathiane with benzyl
bromide. We have found that condensation of phenylacetaldehyde with
3-mercapto-1-propanol under standard acid-catalyzed Dean-Stark distil-
lation conditions gives 11 in good yield.
(8) Hill, C. M.; Simmons, D. E.; Hill, M. E. J. Am. Chem. Soc. 1955, 77, 3889.
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
631

Scheme 3. Initial route and the first radical addition based route to acid 9
however, whilst a number of methods yielded the desired
product, the reaction was usually accompanied by the
formation of the symmetrical ketone 10 as a major impurity.
Clearly, this impurity arises from condensation of 2 mol-
ecules of a derivative of the acid 9, such as the active ester
or derived ketene, followed presumably by a decarboxylation.
Another troublesome impurity was the N,O-diacyl derivative
formed by acylation of both the amino and phenol hydroxy
functions, observed in most of the coupling methods
screened. There were a number of other impurities observed,
some specific to the coupling method employed, but in all
cases after workup and during isolation, the amide proved
reluctant to solidify, and isolated yields and purities were
low.
The reduction stage proved no less problematic, even
when the input amide had been purified by chromatography.
Borane-mediated reactions were characterized by over-
reduction, involving the benzothiazolone carbonyl function,
resulting in several impurities which were difficult to remove
and appeared to inhibit crystallization of the hydrochloride
1 after workup. Preparative chromatography was required
to obtain pure material and crystallization of the hydrochlo-
ride salt from methanol was then facile but in an overall
yield of only 20% from 2. Although elimination of the
chromatography may have been possible with extended
optimization studies to give material of acceptable quality,
we took the view that the low yields and volume efficiencies
associated with these stages would render long-term through-
put requirements unattainable using this route.
sibenadet hydrochloride from 2 was required, and a straight-
forward disconnection across the carbon-nitrogen bond led
us to consider a direct N-alkylation of the free amine 2b
(Scheme 4). Since the side chain of sibenadet contains a
â-amino sulfone moiety, an obvious choice of electrophile
for the alkylation of 2b is the vinyl sulfone 15 which should
act as acceptor in a conjugate addition reaction.9 It was our
hope that an electron-withdrawing inductive effect associated
with the beta disposed sulfone group, once appended, would
counteract the usual increase in nucleophilicity associated
with secondary amine products in such alkylations and
minimize formation of tertiary amine byproducts that often
hampers this method as a preparative technique.10
Initial model studies using phenyl vinyl sulfone11 as
electrophile for amine 2b were encouraging, and a rapid entry
to the vinyl sulfone 15 was sought in order to explore this
route. We took advantage of the existing laboratory supplies
of acid 9 as a precursor (Scheme 5) by reducing to the
alcohol, converting this to the chloride in situ and eliminating
(9) (a) Teyssot, M.-L.; Fayolle, M.; Philouze, P.; Dupuy, C. Eur. J. Org. Chem.
2003, 54. (b) Bashford, K. E.; Cooper, A. L.; Kane, P. D.; Moody, C. J.;
Muthusamy, S.; Swann, E. J. Chem. Soc., Perkin Trans. 1 2002, 14, 1672.
(c) Evans, B. E.; Lundell, G. F.; Gilbert, K. F.; Bock, M. G.; Rittle, K. E.;
Carroll, L. A.; Williams, P. D.; Pawluczyk, J. M.; Leighton, J. L.; Young,
M. B.; Erb, J. M.; Honns, D. W.; Gould, N. P.; DiPardo, R. M.; Hoffman,
J. B.; Perlow, D. S.; Whitter, W. L.; Veber, D. F.; Pettibone, D. J.;
Clineschmidt, B. V.; Anderson, P. S.; Freidinger, R. M. J. Med. Chem.
1993, 36, 3993. (d) Schoeberl, A.; Biedermann, M. Justus Liebigs Ann.
Chem. 1968, 716, 37. (e) Friedman, M.; Romersberger, J. A. J. Org. Chem.
1968, 33, 154.
(10) Srivastava, S. K.; Chauhan, P. M. S.; Bhaduri, A. P. Synth. Commun. 1999,
29, 2085 and references therein.
(11) Rosenmund, P.; Brandt, B.; Flecker, P.; Hoffmann, E. Liebigs Ann. Chem.
1990, 857.
Route Selection for Synthesis of Sibenadet Hydro-
chloride - the Alkylation Route: An alternative route to
632
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

Scheme 4. Alkylation route to 1: addition of 2b to a vinyl
sulfone
the expected tertiary amine and, interestingly, the S-alkylated
isomer (∼10 mol % in total). The tertiary amine was
characterized initially by HPLC-MS and subsequently con-
firmed by NMR experiments run on a sample isolated by
chromatography that demonstrated the existence of two
equivalent side chain units. The structure of the S-alkylated
isomer was originally misassigned as the N,N′-bis-alkylated
1
derivative but H, 13C, and several 2-D NMR experiments
confirmed that the thiazolone ring had been replaced by an
oxazolone ring and the sulfur atom had become alkylated.
Nevertheless, the levels of these impurities were reduced
upon isolation of 1 in an initial crystallization, and attempted
purification by recrystallization indicated that methanol
appeared to be a suitable solvent, yielding material that
should be acceptable from a quality viewpoint.
These results demonstrated some major advantages of this
method over the amide route, cutting out a synthetic step
and at the same time removing the need for preparative
chromatography. We believed that this approach should
deliver a long-term manufacturing process and set about
process research towards an improved synthesis of vinyl
sulfone 15, subsequent scale-up work, and further develop-
ment and scale-up of the alkylation process to give 1.
Route Selection for Synthesis of Vinyl Sulfone 15: The
Final Synthetic Route for 1. The overall method employed
to prepare vinyl sulfone 15 (Schemes 3 and 5) was not
particularly efficient, since it incorporated what appeared to
be an unnecessary reduction of acid 9. Therefore, a strategy
based upon analogous radical addition chemistry early on
using 2-mercaptoethanol instead of mercaptoacetic acid was
examined (Scheme 6).
Scheme 5. Initial route used to prepare vinyl sulfone 15
This thiyl radical addition reaction worked in a similar
mode to its predecessor, resulting in good conversion to give
thioether 16 along with around 10 mol % of the regioisomeric
addition product 16a. The radical reaction could be initiated
by light, air, or chemical initiator, and azobisisobutyronitrile
was found to be most suitable. The thioether could be
oxidized smoothly to the sulfone 17 using hydrogen peroxide
and acetonitrile under basic conditions.12 We envisaged that
conversion of 17 to the vinyl sulfone 15 should be possible
by converting the hydroxy group into a good nucleofuge for
a â-elimination reaction. A variety of derivatives were made,
and it was found that under the right conditions esters were
reactive enough toward â-elimination to allow facile prepara-
tion of 15.13 An example, the benzoate ester 18, was also
found to be highly crystalline, and this greatly enhanced the
route as a potential manufacturing process, since all the
preceding intermediates were oils. Suitable conditions for
the elimination included DBU in ethyl acetate at ambient
temperature and, most conveniently, triethylamine in indus-
trial methylated spirit (IMS; ethanol denatured with ∼4%
methanol) at reflux. The crude benzoate 18 was derived from
2-phenyl ethanol in what was essentially a telescoped
process, and once it was established that 18 could be
crystallized in a highly pure state, free from the regioisomer
HCl to give 15 as a low melting solid. Initial reactions of
15 with 2 in the presence of triethylamine in both DMF and
methanol demonstrated the feasibility of the approach, with
reasonable conversion to 1a, and in the case of the methanol
reactions, product of good quality could be isolated as the
filterable solid hydrochloride 1 by acidification with hydro-
chloric acid in 60-80% yield. A number of impurities were
observed in the crude reaction mixtures, including two
byproducts shown in Scheme 4 derived from over-alkylation,
(12) Ando, T.; Cork, D. G.; Kimura, T. Chem. Lett. 1986, 5, 665.
(13) Eyley, S. (AstraZeneca AB). WO 00/50413 A1.
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
633

Scheme 6. Final route to 15 and conditions used for
large-scale production of 18
Scheme 7. Production process used for manufacture of 1
from 2
and Phase III clinical trials and carcinogenicity studies.
Commercial Process
Benzothiazolone Amine Hydrochloride 2. The nine-
stage process for manufacture of benzothiazolone hydro-
chloride 2 outlined in Scheme 1 was transferred to external
supply manufacturers for further development and scale-up
to production where process improvements were introduced.
In addition to the enhanced nitration procedure already
discussed, another significant change was made in the
deprotection of the benzoyl thiourea 5, where alternative
conditions were employed, involving use of catalytic sodium
methoxide in methanol with an acetic acid quench. This
proved more robust than the hydroxide process, which
required tight control in the acidifying quench, and furnished
material at least as good if not better in terms of quality. It
was later established that either set of conditions could be
used at this processing stage without giving rise to significant
differences in impurity profile at intermediate 2 or at the
drug substance itself.
At this stage, preparation was underway for validation of
the processing stages downstream of intermediate 4, our
proposed registered starting material, and appropriate process
ranges for the key parameters were established along with
suitable specifications for the intermediates. As part of this,
key batch data were taken into account and any anomalous
runs were carefully considered. For the deprotection stage
yielding 2a, investigation of a failed batch showed that the
derived from 16a, we felt that this route was an appropriate
choice as the final synthetic sequence.
Process Development: The Final Manufacturing Pro-
cess. Having demonstrated that elimination of benzoate from
18 to give the vinyl sulfone 15 could be achieved by heating
with triethylamine in IMS, the same solvent type as that used
for the subsequent alkylation of 2b, it seemed logical to
attempt both these transformations in the same reaction
vessel. This would offer a reduction in the number of
processing stages and obviate the need to isolate 15 (Scheme
7). This was reduced to practice in the laboratory by mixing
stoichiometric quantities of 18 and 2 in ethanol, adding
triethylamine, heating to reflux for several hours, cooling,
and acidifying with hydrochloric acid. Crude 1 was found
to crystallize out in a reproducible process. The crude
material obtained was converted to the pure drug substance
by means of a single recystallization from methanol giving
material with an acceptable impurity profile (no significant
contamination by benzoic acid was observed). This repre-
sented a major improvement, and the process, now deemed
suitable for long-term supply of 1, was transferred to the
pilot plant and used to supply drug substance for Phase II
634
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

Scheme 8. Intermediates involved in the HBr mediated
deprotection of 8 to give 2a
Figure 3. Rate constants k4-k7 as a function of HBr concen-
tration.
strength of the hydrobromic acid solution was important. A
kinetic study revealed that the reaction pathway was related
to the acid concentration, and the mechanistic pathway
outlined in Scheme 8 was elucidated, where k1-7 represent
rate constants.
Figure 4. Molar concentration profiles of intermediate 21
calculated from rate constants derived from fitting experimental
data.
Isolation and analysis of the product mixtures from these
small-scale reactions were not carried out; therefore, these
data do not support a specific lower limit for HBr concentra-
tion. However, it is worth noting that a concentration of 5
× 10-4 M for 21 at the end of the reaction would lead to at
least 0.26% w/w of this impurity in the batch of 2 (this
assumes proportionate cocrystallization of 21 with 2a and
no reduction in the recrystallization stages for conversion
of 2a to 2, which is consistent with data from several
laboratory and pilot plant batches). It had been established
that 21 was converted to the corresponding methoxy-
derivative of 1 in the final stages of drug substance
manufacture, an impurity that we wished to control. These
results led to the introduction of drying of precursor 8,
isolated from an aqueous quench, and the incorporation of a
lower limit for HBr concentration (46% w/w) as a key
parameter for robustness.
Another important operational parameter was the type of
filter used to remove the charcoal from the recrystallizations
during conversion of 2a into 2, since physical separation of
the highly insoluble dihydrochloride impurity Bis-2 occurred
during this filtration. It was important to control the level of
this impurity at this stage since it was found to react, not
unexpectedly, in the downstream alkylation chemistry to give
derivatives that would contaminate the final product 1. Using
filters that were rated capable of removing particulates down
to 1 micron reduced the level of Bis-2 present in 2 to within
The first step is a simple rapid displacement of chloride
by bromide (the 2-bromo benzothiazole 19 could be isolated
from low conversion runs), and this in turn undergoes fast
hydrolysis to the benzothiazolone 20. The deprotection
pathway then diverges through two intermediates 21 and 22,
each of which is converted to the final product 2a.
Only those steps whose products have been detected in
time courses of the reaction at reflux temperatures were
included in this analysis (i.e., downstream from 20). The
concentration of HBr was varied between 42 and 57% w/w,
at reflux temperature, using purified 20 as starting material.
Samples were removed at specified time-points and analyzed
by HPLC to determine molar concentrations. The kinetic
parameters, based on the above reaction scheme describing
pseudo-first-order reactions, were determined by fitting
experimental data using RATE in BatchCAD.
Variation of rate constants with HBr concentration
revealed that the rate of the deacetylation was essentially
invariant over the acid concentration range studied. However,
the rate of the demethylation stages showed a strong
dependency on acid concentration (Figure 3).
The profile for the intermediate 21 is therefore sensitive
to the concentration of HBr as depicted in Figure 4, derived
from calculations based on the fitted rate constants. This
indicates that, at lower HBr concentrations, the level of this
compound peaks at much higher values and persists longer
throughout the reaction.
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
635

acceptable specified limits.
sibenadet hydrochloride could exist in more than one
polymorphic form. Two forms relevant to the manufacturing
process were identified, characterized, and designated Form
I and Form II. We were able to demonstrate that Form I
was thermodynamically more stable and that Form II was
meta-stable with respect to Form I and was converted to
Form I in differential scanning calorimetry recycling experi-
ments. This interconversion also translated to the recrystal-
lization process where we found that batches contaminated
with Form II could be converted to Form I on recrystalli-
zation. However, sometimes more than one recrystallization
was required to effect complete conversion. Since we
believed that the ideal process for isolation and purification
of 1 should be limited to a crystallization of the crude
followed by a single recrystallization, we set about optimiza-
tion with respect to control of the polymorphic form, without
compromising the impurity profile.
In an earlier process for recrystallization of 1, a distillation
step was employed to concentrate the methanol solution prior
to final recrystallization. It was observed that this distillation
would often lead to the formation of a crust of 1 that formed
on the surfaces of the hot reaction vessel during solvent
removal. Batches prepared in this way were comprised of
mixtures of Form I and Form II with Form II often
predominating. Further laboratory studies showed that the
majority of recrystallizations where deliberate crust formation
was allowed to occur yielded mixtures of the two forms. In
addition, an extensive series of laboratory recrystallizations
with seeding and low cooling rate (of the order of 0.1 °C
min-1) gave predominantly Form I material. Laboratory
experiments using a turbidiometric method indicated that 1
has a very wide meta-stable zone width in the solubility curve
for this recrystallization system. Since this implied that a
state of high supersaturation was likely to be long-lived,
especially with slow cooling rates, we adopted seeding
routinely for all scale runs in an attempt to avoid uncontrolled
spontaneous crystallization. Unfortunately, the corrosive
nature of the recrystallization system precluded on-line
particle size monitoring with the equipment available at the
time, and so we were unable to determine unequivocally
whether seeding was indeed controlling the crystallization
through secondary nucleation.
Ultimately the manufacturing processes downstream of
4 were fully validated at two commercial suppliers, with an
output batch size of approximately 70 kg for 2, and over 1
tonne of material was produced using the validated process.
Although these process changes were introduced after the
batches of 2 used for Phase III drug substance manufacture
were produced, we were able to demonstrate equivalence at
intermediate 2 and also later confirmed that the drug
substance made from these supplies of 2 was also equivalent
to the Phase III drug substance in terms of both impurity
profile and physical form.
Benzoate Ester 18. The synthesis of allyl ether 14 from
2-phenyl ethanol under phase transfer conditions with
portionwise addition of tetrabutylammonium hydrogen sul-
fate catalyst proceeded smoothly and was scaled up to a 105
kg batch size (input of 2-phenyl ethanol). After an aqueous
wash and a distillation, the product was isolated as a toluene
concentrate. The next three synthetic transformations were
performed on this concentrate in a telescoped sequence. After
the radical addition, again run in toluene, the product solution
was submitted to the oxidation process. Here, our outsourcing
partner introduced a notable modification, involving a
tungstate-catalyzed hydrogen peroxide oxidation14 in a two-
phase aqueous toluene solvent system. This gave a much
improved volume efficiency compared to the original aceto-
nitrile-mediated process. In addition, since the organic solvent
was common to the previous stage, a simple aqueous workup
and azeodrying protocol could be employed giving the
sulfone alcohol 17 as a toluene concentrate. This was
subjected to acylation using benzoyl chloride in isopropyl
acetate as cosolvent, and after a solvent swap, the crude
benzoate 18 crystallized upon seeding. A routine recrystal-
lization gave highly pure 18 in around 70% overall yield
from 2-phenyl ethanol. This process was fully validated and
provided 230 kg output batch sizes for 18. Over 1 tonne of
material was produced using the validated process.
Sibenadet Hydrochloride 1. The process defined for
manufacture of the batches for Phase III clinical trials
(Scheme 7) was transferred successfully to the production
facility with a 3-fold increase in scale. As part of this, the
following process improvements were made: (i) seeding was
introduced for crystallization of the crude drug substance,
and the wash solvent was changed from two portions of
2-propanol to a single methanol wash, improving filterability
considerably; (ii) the pure drug substance was also washed
with methanol instead of 2-propanol, again improving
filterability and facilitating subsequent drying. These changes
were predicated on technical trial batches run in the pilot
plant prior to full technology transfer to production, and on
laboratory based process range work, and confirmation was
obtained during commissioning in production that the
changes had no significant impact of quality whilst improving
the overall cycle time.
During manufacture of the drug substance for Phase III
clinical trials, based on these empirical data, the following
control measures were adopted: transfer of the hot solution
from dissolution vessel to recrystallization vessel (via a
polishing filter) under conditions that avoided evaporation
and adventitious crystallization, addition of seed (minimum
of 3.1% w/w) between 50 and 60 °C, and a cooling rate of
0.1 °C min-1. These parameters were formally transferred
to the commercial manufacturing facility during commis-
sioning. All scale runs using this recrystallization protocol
afforded Form I material. The final two stages giving crude
and pure sibenadet hydrochloride were successfully validated
at a 40 kg batch output scale.
Control of Physical Form. During preparation of drug
substance for Phase II clinical trials, it was discovered that
Conclusions
(14) (a) Stec, Z.; Zawadiak, J.; Skibinski, A.; Pastuch, G. Polish J. Chem. 1996,
70, 1121. (b) Bortolini, O.; Di Furia, F.; Modena, G.; Seraglia, R. J. Org.
Chem. 1985, 50, 2688.
A process for manufacture of initial launch quantities of
sibenadet hydrochloride 1 has been developed based on a
636
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

convergent synthesis involving key intermediates 2 and 18.
It was estimated that the validated batch scale would supply
sufficient drug substance annually for three years postlaunch
with further process efficiency improvements. Projected peak
annual demand four years postlaunch of around 3 tonnes
could, if necessary, be met through increasing the batch scale.
acetic acid (460 kg), N-[4-methoxyphenyl)ethyl]acetamide
(115 kg, 595 mol), acetic anhydride (102 kg, 999 mol), and
98% sulfuric acid (11 kg, 110 mol). To this mixture was
added 98% nitric acid (56 kg, 871 mol) in 5 kg aliquots
maintaining a reaction temperature of 15-20 °C during the
addition. The reaction mixture was stirred for 1 h and then
added to water (1150 L), and the mixture stirred for 16 h.
20% Sodium hydroxide solution (2200 L) was added to the
stirred mixture over 6 h keeping the temperature below 25
°C, and the product was collected by filtration, washed with
water, and dried in a vacuum tray oven. This gave N-[3-
nitromethoxyphenyl)ethyl]acetamide (99.1 kg, 70% yield
corrected for moisture) as a solid that contained approxi-
mately 5% w/w moisture which was suitable for processing
in the next stage. An analytically pure sample was obtained
by recrystallization from two volumes of ethyl acetate (53%
recovery) giving yellow crystals, mp 88-89 °C, GC purity
Experimental Section
General. Melting points were recorded on a capillary
melting point apparatus and are uncorrected. 1H NMR spectra
were recorded at 300, 360, or 400 MHz in CDCl3 or DMSO-
d6. 13C NMR spectra were recorded at 75, 100, or 125 MHz
in DMSO-d6 (resonances coincident with those of the solvent
are not quoted). Mass spectra were recorded on a variety of
instruments using electron impact (EI), attached proton
chemical ionisation (APCI), or fast atom bombardment
(FAB) ionisation techniques.
1
For compounds 1, 2, 4-8, 14, 17, and 18, the procedures
described represent the manufacturing scale processes that
have been fully validated.
99.1%; MS (EI) m/z 238 (M+ + H); H NMR (360 MHz,
CDCl3) δ 7.67 (d, J ) 2.2 Hz, 1H), 7.39 (dd, J ) 8.6, 2.2
Hz, 1H), 7.04 (d, J ) 8.6 Hz, 1H), 5.66 (br s, 1H), 3.94 (s,
3H), 3.49 (q, J ) 7.1 Hz, 2H), 2.82 (t, J ) 7.1 Hz, 2H),
1.96 (s, 3H). Anal. Calcd for C11H14N2O4: C, 55.46; H, 5.92;
N, 11.76. Found: C, 55.56; H, 6.05; N, 11.81.
N-[4-Methoxyphenyl)ethyl]acetamide.15 A suitable reac-
tion vessel was charged with acetic anhydride (103.5 kg,
1014 mol), and 2-(4-methoxyphenyl)ethylamine (143.4 kg,
948 mol) was added from a head tank holding the temper-
ature at less than 30 °C (range 20 to 30 °C). After the
addition was complete, the reaction mixture was stirred for
1 h. A suitable quench vessel was charged with 32% sodium
hydroxide (352 kg, 2816 mol) and water (399 L). The
reaction mixture was added to the sodium hydroxide solution
keeping the temperature at less than 20 °C. The quench slurry
was cooled to 0-5 °C, and the precipitated product was
isolated on a centrifuge and washed with water (497 L). The
product was obtained as a water-wet cake and used directly
in the next stage.
N-[2-(3-Amino-4-methoxyphenyl)ethyl]acetamide, Hy-
drohloride (4). A suitable vessel was charged with 2-pro-
panol (2244 L) and N-[3-nitromethoxyphenyl)ethyl]aceta-
mide (375 kg, 1574 mol), and then a slurry of 5% palladium
on charcoal (31 kg, 50% water wet) in 2-propanol (100 L)
was added and washed in with 2-propanol (100 L). Hydrogen
was introduced to the stirred mixture allowing the temper-
ature to rise to 62-68 °C and the pressure to rise to 2.5-
3.0 bar. When the theoretical charge of hydrogen has been
added (∼9.5 kg), the reaction mixture was filtered at ∼65
°C into a suitable vessel to remove the catalyst which was
washed on the filter with 2-propanol (298 L). To the
combined, stirred filtrate was added 33% hydrochloric acid
(261 kg, 2363 mol) at 17-23 °C over ∼1 h. The mixture
was stirred for at least 4 h at 17-23 °C, the precipitated
product was isolated on a centrifuge in four loads, and each
load was washed with 2-propanol (3 × 67 L). The damp
product was dried under vacuum at up to 80 °C to give 4
(310 kg, 80%) as an off-white solid, HPLC purity 99.2%;
N-[3-Nitro-4-methoxyphenyl)ethyl]acetamide. Method
A. A suitable dissolution vessel was charged with acetic acid
(87 kg), N-[4-methoxyphenyl)ethyl]acetamide (29.0 kg dry
weight equivalent, 150 mol) was added, and the contents of
the vessel were warmed to 30-35 °C. The resulting solution
was cooled to 15-25 °C and charged to a head tank
connected to a suitable reaction vessel. Concentrated nitric
acid (70%, 135 kg, 1500 mol) was charged to the reaction
vessel, and the solution of acetamide was added from the
head tank keeping the temperature in the range 15-25 °C.
After the addition was complete, the reaction mixture was
stirred for 30 min at 15-25 °C and cooled to 0-5 °C and
then added to a solution of 20% sodium hydroxide (591 kg,
2955 mol) in a suitable quench vessel keeping the temper-
ature in the range 0-20 °C. The quench slurry was cooled
to 0-5 °C, and the precipitated product was isolated on a
centrifuge and washed with water (123 L). The product was
obtained as a water-wet cake. Yield 17.9 kg, 50% from 2-(4-
methoxyphenyl)ethylamine (corrected for moisture).
Method B (Employed Using N-[4-Methoxyphenyl)-
ethyl]acetamide That Had Been Dried to <2% w/w
Moisture). A suitable dissolution vessel was charged with
1
MS (APCI +ve) m/z 209 (M+ + H); H NMR (300 MHz,
DMSO-d6) δ 10.06 (br m, 3H), 8.02 (br t, J ) 5.4 Hz, 1H),
7.30 (d, J ) 1.8 Hz, 1H), 7.21 (dd, J ) 8.4, 1.8 Hz, 1H),
7.18 (d, J ) 8.4 Hz, 1H), 3.86 (s, 3H), 3.24-3.16 (m, 2H),
2.65 (t, J ) 7.2 Hz, 2H), 1.79 (s, 3H); 13C NMR (75 MHz,
DMSO-d6) δ 169.1, 150.6, 132.1, 129.1, 123.9, 120.4, 112.4,
56.1, 34.1, 22.6. Anal. Calcd for C11H16N2O2.HCl: C, 53.99;
H, 7.00; N, 11.45. Found: C, 53.72; H, 6.95; N, 11.34.
N-{2-(4-Methoxy-3-[3-(1-phenylmethanoyl)thioureido]-
phenyl)ethyl}acetamide (5). Reaction vessel A was charged
with water (2380 L), 4 (485 kg, 1982 mol), and methyl
isobutyl ketone (1905 kg). 30% Sodium hydroxide was added
at 15-25 °C until the pH of the aqueous layer was >12
(approximately 340 kg). The contents of the reactor were
heated to 50-55 °C, and the two layers that had formed
were allowed to separate. The organic layer was retained,
(15) (a) Yamada, S.; Fujii, T.; Takagi, K.; Gomi, Y.; Matsushita, S. Chem. Pharm.
Bull. 1963, 11, 258. (b) Walpole, G. S. J. Chem. Soc. 1910, 97, 941.
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
637

and the aqueous layer was further extracted with methyl
isobutyl ketone (1575 kg) at 50-55 °C. The organic layers
were combined and retained in reaction vessel A at 50-55
°C. Reaction vessel B was charged with methyl isobutyl
ketone (1540 kg), ammonium thiocyanate (182 kg, 2391
mol), and benzoyl chloride (311 kg, 2212 mol), maintaining
the temperature at 35-40 °C, stirred for 1 h, and then cooled
to 15-25 °C. The resulting solution of benzoyl isothiocy-
anate was transferred, via a filter to remove ammonium
chloride, into reaction vessel A (containing the solution of
free base of 4) at 50-55 °C. The filter cake was washed
with methyl isobutyl ketone (230 kg), and this washing added
to reaction vessel A. The mixture was stirred for a minimum
of 2 h at 50-55 °C and cooled to 15-25 °C, and the product
was isolated by filtration and washed with methyl isobutyl
ketone (645 kg) and water (1320 kg). The damp product was
dried under vacuum at up to 65 °C to give 5 (663 kg, 90%)
as an off-white solid, HPLC purity >97.5%. An analytically
pure sample was obtained by slurrying in 20 volumes of
refluxing acetonitrile, cooling, and filtering to afford a white
solid (96% recovery), HPLC 98.9%; MS (APCI +ve) m/z
372 (M+ + H); IR (neat, cm-1) 3300, 1640, 1560, 1523; 1H
NMR (300 MHz, DMSO-d6) δ 13.00 (s, 1H), 11.55 (s, 1H),
8.50 (s, 1H), 8.00-7.97 (m, 2H), 7.89 (t, J ) 5.4 Hz, 1H),
7.70-7.64 (m, 1H), 7.57-7.51 (m, 2H), 7.07 (s, 2H), 3.87
(s, 3H), 3.27-3.20 (m, 2H), 2.67 (t, J ) 7.2 Hz, 2H), 1.80
(s, 3H); 13C NMR (75 MHz, DMSO-d6) δ 177.7, 169.0,
168.4, 149.0, 132.0, 130.9, 128.7, 128.4, 126.6, 126.5, 123.4,
111.2, 56.1, 34.6, 22.7. Anal. Calcd for C19H21N3O3S: C,
61.43; H, 5.70; N, 11.31. Found: C, 61.36; H, 5.68; N, 11.29.
N-[2-(4-Methoxy-3-thioureidophenyl)ethyl]acetamide
(6). A suitable vessel was charged with methanol (2160 kg)
and 5 (550 kg, 1481 mol), and the mixture was heated to
40-45 °C. Sodium methoxide in methanol (21%, 83 kg, 323
mol) was added, and the reaction mixture was stirred at 40-
45 °C for 45-60 min. Acetic acid (22 kg, 366 mol) was
added, the vessel contents were cooled to 15-20 °C and
stirred for 30 min, and the precipitated product was collected
by filtration and washed with methanol (470 kg) followed
by water (590 kg). The damp product was dried under
vacuum at up to 80 °C to give 6 (399 kg, 90%) as an off-
white solid, HPLC purity >99%; MS (APCI +ve) m/z 268
(M+ + H); IR (neat, cm-1) 3435, 1615, 1523; 1H NMR (300
MHz, DMSO-d6) δ 8.97 (s, 1H), 7.83 (t, J ) 5.4 Hz, 1H),
7.80-7.00 (br m, 3H), 7.00-6.93 (m, 2H), 3.78 (s, 3H),
3.24-3.17 (m, 2H), 2.63-2.57 (m, 2H), 1.77 (s, 3H); 13C
NMR (100 MHz, DMSO-d6) δ 181.2, 169.0, 150.3, 131.0,
127.0, 125.9, 111.4, 55.6, 34.5, 22.6. Anal. Calcd for
C12H17N3O2S: C, 53.91; H, 6.40; N, 15.71. Found: C, 53.67;
H, 6.39; N, 15.52.
30 min. The reaction mixture was stirred for 30-40 min at
5-10 °C, warmed to 45-50 °C, and stirred for a further
60-70 min and then cooled to 20-25 °C. The mixture was
added to a cold 21% sodium hydroxide solution (9600 kg,
prechilled to 5-10 °C), keeping the temperature below 30
°C during the addition. The temperature of the mixture was
adjusted to 20-25 °C, and the product was isolated by
filtration and washed with water until the pH of the washings
was in the range 6-8 (approximately 2900 kg). The damp
product was dried under vacuum at up to 75 °C to give 7
(265 kg, 92%) as an off-white solid, HPLC purity >97%;
MS (APCI +ve) m/z 266 (M+ + H); IR (neat, cm-1) 3405,
3317, 1661, 1629, 1540, 1501; 1H NMR (400 MHz, DMSO-
d6) δ 7.94-7.91 (m, 1H), 7.42 (s, 2H), 6.80-6.76 (m, 2H),
3.81 (s, 3H), 3.28 (q, J ) 7.2 Hz, 2H), 2.69 (t, J ) 7.2 Hz,
2H), 1.78 (s, 3H); 13C NMR (100 MHz, DMSO-d6) δ 169.1,
164.9, 148.5, 141.9, 132.0, 124.4, 121.2, 108.6, 55.8, 35.0,
22.7. Anal. Calcd for C12H15N3O2S: C, 54.32; H, 5.70; N,
15.84. Found: C, 54.28; H, 5.68; N, 15.87.
N-[2-(2-Chloro-4-methoxybenzothiazol-7-yl)ethyl]acet-
amide (8). A suitable vessel was charged with copper(I)
chloride (29 kg, 293 mol), copper(II) chloride dihydrate (50
kg, 293 mol), ethanol (8 kg), and 33% hydrochloric acid
(1991 kg, 1800 mol). 6 (155 kg, 584 mol) was added
portionwise, with the temperature kept below 28 °C during
the addition. The resulting suspension was cooled to 2-8
°C, and 28% sodium nitrite solution (432 kg, 1753 mol) was
added over 1.75-2.0 h keeping the temperature at 2-8 °C
and held within this temperature range for a further 90 min.
The reaction mixture was added to a mixture of water (6098
L) and methyl isobutyl ketone (430 kg) at 20-25 °C, and
the resulting mixture was stirred for 17.5 h at 20-25 °C.
The precipitated product was isolated by filtration and
washed with water (1860 L). The damp product was dried
under vacuum at up to 50 °C to give 8 (107 kg, 64%) as a
red-brown solid, HPLC purity >94%. An analytically pure
sample was obtained by slurrying the crude material in 16
volumes of ethyl acetate, refluxing for 3 h, and filtering hot
to remove Bis-8. Addition of charcoal to the hot solution
(50% w/w relative to crude 8), refluxing for a further 1 h,
hot filtration, and cooling overnight yielded 8 as yellow-
orange crystals (51% recovery), HPLC purity 99.7%; MS
(APCI +ve) m/z 285 and 287 (M+ + H), 251 (M - Cl); IR
(neat, cm-1) 3364, 1673; 1H NMR (300 MHz, DMSO-d6) δ
7.97-7.93 (m, 1H), 7.29 (d, J ) 8.1 Hz, 1H), 7.08 (d, J )
8.1 Hz, 1H), 3.93 (s, 3H), 3.37-3.29 (m, 2H), 2.83 (t, J )
7.2 Hz, 2H), 1.77 (s, 3H); 13C NMR (125 MHz, DMSO-d6)
δ 169.2, 151.2, 149.5, 140.0, 137.5, 126.9, 124.7, 108.8, 56.0,
38.3, 34.5, 22.6. Anal. Calcd for C12H13ClN2O2S: C, 50.61;
H, 4.60; N, 9.84. Found: C, 50.57; H, 4.59; N, 9.78.
7-(2-Aminoethyl)-4-hydroxy-1,3-benzothiazol-2(3H)-
one, Hydrochloride (2). A suitable vessel was charged with
48% hydrobromic acid (2805 kg) and 8 (170 kg, 597 mol).
The mixture was heated to 115 °C and then held at reflux
(115-125 °C) for 7.0-7.5 h before cooling to 0-10 °C and
stirring for about 1 h. The hydrobromide salt (2a) was
isolated and washed with 2-propanol (total 268 kg). The
damp hydrobromide salt (∼160 kg damp weight) was
N-[2-(2-Amino-4-methoxybenzothiazol-7-yl)ethyl]acet-
amide (7). A suitable vessel was charged with methane-
sulfonic acid (1825 kg) and acetic acid (367 kg), and 6 (290
kg, 1085 mol) was added portionwise to the stirred mixture
keeping the temperature of the mixture below 30 °C. The
resulting mixture was cooled to 5-10 °C, and a freshly
prepared solution of N-bromosuccinimide (184 kg, 1034 mol)
in methanesulfonic acid (590 kg) was added at 5-10 °C over
638
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

charged to a suitable vessel containing water (1632 kg),
charcoal (Norit SX1-G carbon, 14 kg), and filter aid (Vitacel
FAC200, 7 kg). The mixture was warmed to 65-75 °C and
held for at least 30 min and then filtered hot through a
polishing filter capable of removing particles down to at least
1 micron in size. The filtrate was held at 65-75 °C, and the
filter and pipe work were rinsed through with hot water at
65-75 °C (68 kg). To the stirred filtrate was added 33%
hydrochloric acid (302 kg, 2733 mol) at 60-70 °C, and then
the mixture was cooled to 5-15 °C. The slurry was held at
5-15 °C for at least 30 min, and the product was isolated
by filtration and washed with 2-propanol (total 217 kg). This
material (∼135 kg damp weight) was recrystallized once
more in the same way using water (1233 kg), charcoal (10
kg), filter aid (7 kg), wash water (51 kg), and 33%
hydrochloric acid (225 kg, 2037 mol). The recrystallized
material was isolated by filtration, washed with 2-propanol
(total 170 kg), and dried under vacuum at up to 70 °C to
give 2 (93 kg, 63%), as a light yellow crystalline solid, HPLC
purity >98.5%; MS (APCI +ve) m/z 211 (M+ + H); IR
(neat, cm-1) 3127, 3062, 1665; 1H NMR (400 MHz, DMSO-
d6) δ 11.80 (br s, 1H), 10.19 (br s, 1H), 8.14 (br s, 3H),
6.87 (d, J ) 8.2 Hz, 1H), 6.79 (d, J ) 8.2 Hz, 1H), 2.98 (br
m, 2H), 2.79 (t, J ) 7.8 Hz, 2H); 13C NMR (100 MHz,
DMSO-d6) δ 169.1, 141.8, 124.7, 123.6, 123.0, 121.1, 112.8,
38.2, 31.1. Anal. Calcd for C9H10N2O2S.HCl: C, 43.81; H,
4.49; N, 11.35; Cl, 13.00; S, 14.37. Found: C, 44.22; H,
4.53; N, 11.47; Cl, 13.37, S, 14.63.
[2-(2-Propenyloxy)ethyl]benzene (14). A suitable vessel
was charged with toluene (470 L), stirring was started, and
sodium hydroxide (micropearls, 103 kg, 2575 mol) followed
by tetrabutylammonium hydrogen sulfate (TBAH) (3.15 kg)
was added. 2-Phenylethanol (105 kg, 860 mol) and allyl
bromide (135 kg, 1116 mol) were added and rinsed in with
toluene (30 L). The vessel was heated to 65-75 °C, and the
contents were maintained at 65-75 °C for 2-4 h with
efficient stirring. A solution of TBAH (2.10 kg) in water
(5.3 L) was charged to the vessel, and stirring continued at
65-75 °C for 1-2 h. A further portion of TBAH (1.05 kg)
in water (2.6 L) was added to the vessel, and stirring
continued at 65-75 °C for up to 16 h. The mixture was
cooled to 35-40 °C, water (210 L) was added, and the
resulting mixture stirred for 15 min. Stirring was stopped,
and the phases were allowed to separate at 35-40 °C. The
aqueous phase was run to waste, and the organic phase was
washed with water (55 L). This wash was repeated, and the
organic phase was separated and concentrated by distillation
of the solvent (approximately 530 L of solvent was removed;
if necessary distillation was continued until allyl bromide
content was e0.1% by GC) with a final vessel temperature
of 145 °C. The mixture was cooled to 20 °C and filtered to
give 14 as a toluene concentrate (approximately 95 mol %
14 by GC; equivalent to 135 kg, 97%).
65-70 °C during addition). A solution of AIBN (2.7 kg) in
toluene (140 L) was added at 65-70 °C over 3-5 h and
rinsed in with toluene (20 L). The resulting mixture was
stirred at 65-70 °C for a further 2-3 h and then cooled to
35-45 °C giving a solution of crude thioether 16 used
without further purification. A solution of sodium tungstate
dihydrate (1.4 kg, 4.2 mol) in water (10 L) was added, and
to the resulting stirred mixture was added 35% hydrogen
peroxide (185 L, 2100 mol) at a rate that maintained the
temperature at 40-55 °C during the addition. The resulting
mixture was stirred for 4-6 h at 35-55 °C, stirring was
stopped, and the phases were allowed to separate over 30-
60 min keeping the temperature 35-55 °C. The lower
aqueous layer was separated, extracted with isopropyl acetate
(200 L) by stirring for 15-30 min at 35-55 °C, and the
isopropyl acetate extract combined with the remaining
organic mixture. To the combined organic phases was added
a solution of sodium sulfite (21 kg) in water (200 L), with
the temperature maintained at 35-45 °C during the addition.
The mixture was stirred for 15-30 min at 35-45 °C, stirring
stopped, and the phases were allowed to separate over 30-
60 min. The aqueous layer was run to waste, and the organic
phase was diluted with isopropyl acetate (150 L) and washed
at 40-60 °C with sodium carbonate solution (10% w/v, 7
kg sodium carbonate + 70 L water). This wash was repeated,
and the organic phase was finally washed with sodium
chloride solution (5% w/v, 3.5 kg sodium chloride + 70 L
water) at 40-60 °C. The organic phase was dried by
azeotropic distillation under Dean-Stark conditions collect-
ing approximately 20 L of water with a final vessel
temperature of 92-102 °C. Isopropyl acetate (70 L) was
distilled off, and the reaction mixture was diluted with
isopropyl acetate (200L) and sampled for water content by
Karl-Fisher assay (if water content was >0.3% w/w,
azeotropic distillation was continued). The reaction mixture
was cooled to 15-25 °C and filtered to give a solution of
crude 17 as a toluene/isopropyl acetate solution (yield
approximately 215 kg including regioisomer, approximately
870 kg total weight of solution) used directly in the next
stage.
Benzoic Acid, 2-{[3-(2-phenethyloxy)propyl]sulfonyl}-
ethyl Ester (18). To the toluene/isopropyl acetate solution
of 17 (approximately 870 kg) was added triethylamine (96
kg, 948 mol) and isopropyl acetate (330 L). The mixture
was cooled to 15-30 °C, and benzoyl chloride (134 kg, 953
mol) added over 3-6 h while maintaining the vessel contents
at 15-30 °C during the addition (residual benzoyl chloride
was washed in with isopropyl acetate (50 L)). The resulting
mixture was stirred for 1-2 h at 15-30 °C and then sampled
for conversion by HPLC. If required, a further portion of
triethylamine (8 kg, 79 mol) and benzoyl chloride (11 kg,
78 mol) was added, and the mixture stirred for a further 1-2
h. Water (410 L) was added, and the mixture stirred for 15-
30 min and was then allowed to settle for 30-60 min. The
lower aqueous layer was run to waste, and the organic phase
was washed with a solution of sodium bicarbonate (25 kg
of sodium bicarbonate + 410 L of water) by stirring for 1-2
h and allowing the phases to separate over 30-60 min. The
{[3-(2-Phenethyloxy)propyl]sulfonyl}ethanol (17). To
a solution of mercaptoethanol (72 kg, 921 mol) in toluene
(80 L) at 65-70 °C was added AIBN solid (0.8 kg) in one
portion followed by 14 (toluene concentrate 135 kg weight
equivalent) added over approximately 1 h (mixture held at
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
639

aqueous layer was run to waste, and this wash was repeated.
The organic phase was concentrated by distillation at
atmospheric pressure until the vessel temperature reached
90-95 °C and then further concentrated under progressive
vacuum without allowing the vessel temperature to fall below
60 °C until the vessel temperature reached 90-95 °C at 20-
50 mmHg pressure (a total distillate of approximately 1100
L was collected). The mixture was cooled to 60-70 °C,
warm (approximately 50 °C) 2-propanol (950 L) added, and
the resulting mixture was reheated, if necessary, to 50-55
°C and then cooled to 38-42 °C. Seed crystals of 18 (0.5-
1.5 kg) were added, and the mixture was stirred for 1-2 h
at 35-40 °C and then cooled to 10-20 °C over 2-4 h and
stirred 1-2 h. The crystalline product was isolated by
filtration on a centrifuge, washing with cold (5-15 °C)
2-propanol (2 × 50 L per centrifuge load) to give damp crude
18 (approximately 244 kg dry weight equivalent). A suitable
vessel was charged with damp crude 18 (approximately 244
kg dry weight equivalent) and 2-propanol (980 L), and the
mixture heated to approximately 80 °C to effect dissolution.
The solution was filtered hot into a second vessel, the first
vessel was charged with 2-propanol (120 L) that was then
heated to approximately 80 °C, and this was also passed
through the filter into the second vessel. The resulting
solution in the second vessel was cooled to 42-46 °C and
held at that temperature for at least 1 h. The solution was
then cooled further to 10-20 °C over at least 2 h and held
at 10-20 °C for at least 1 h. The recrystallized product was
collected by pressure filtration, and the filter cake was
washed with prefiltered (0.5 micron cartridge) 2-propanol
(240 L) at 10-20 °C and was dried at 35-45 °C to give
pure 18 (226 kg, 70% from 2-phenylethanol) as a white
crystalline solid, mp 65-66 °C, HPLC purity >99.7%; MS
(APCI +ve) m/z 377 (M+ + H); IR (neat, cm-1) 2938, 1726,
20 °C using a cooling rate of 0.2 °C/min and stirred for at
least 1 h. The crystalline product was collected by filtration,
washed with methanol (105 kg), and dried at up to 55 °C to
give crude 1 (46 kg, 75%) as a light yellow crystalline solid,
HPLC purity >97%.
Recrystallization of 1. Crude 1 (46 kg, 92 mol) was
charged to a dissolution vessel, and methanol (736 kg) was
added with vigorous stirring. The mixture was heated to
reflux (the boiling point was about 65 °C) and held at reflux
for 30 min. Hydrochloric acid (36%, 18.3 kg) was added,
and the mixture was held at reflux for 20 min and then
transferred to a receiving vessel via a polishing filter,
ensuring that crystallization of the product caused by either
overcooling or solvent evaporation was avoided during the
transfer. The vessel contents were allowed to cool to around
55 °C, and seed crystals of 1 (2.28 kg, 4.55 mol) were added.
The mixture was stirred for 90 min at around 55 °C, cooled
to around 8 °C using a cooling rate of 0.1 °C/min, and stirred
for at least 1 h. The solid product was isolated on a centrifuge
in as many loads as necessary, and each load was washed
with methanol (total amount of methanol used for the
washing was 75-120 kg). The product was dried at around
55 °C (approximately 16 h) to give pure 1 (42.3 kg,
equivalent to 40 kg, 87% from crude 1) as a pale yellow
crystalline solid, mp 219-223 °C (dec), HPLC purity >99%;
MS (FAB +ve) m/z 465 (M+ + H); IR (neat, cm-1) 3328,
1
3184, 1655, 1314, 1136, 1117, 606; H NMR (360 MHz,
DMSO-d6) δ 11.81 (s, 1H), 10.19 (s, 1H), 9.43 (br s, 2H),
7.32-7.15 (m, 5H), 6.88 (d, J ) 8.2 Hz, 1H), 6.79 (d, J )
8.2 Hz, 1H), 3.63-3.55 (m, 4H), 3.60 (t, J ) 6.1 Hz, 2H),
3.41-3.30 (br m, 2H), 3.24-3.19 (m, 2H), 3.19-3.11 (br
m, 2H), 2.91-2.85 (m, 2H), 2.81 (t, J ) 7.0 Hz, 2H), 1.92
(m, 2H). 13C NMR (100 MHz, DMSO-d6) δ 169.1, 141.8,
138.9, 128.8, 128.2, 126.0, 124.8, 123.8, 123.0, 120.8, 112.7,
70.9, 67.7, 49.5, 47.8, 46.1, 35.4, 29.8, 21.8. Anal. Calcd
for C22H29N2O5S2‚HCl: C, 52.74; H, 5.83; N, 5.59; S, 12.80.
Found: C, 52.58; H, 6.07; N, 5.75; S, 12.77.
Preparation of Acid 9 from Allyl Ether 14. To a toluene
concentrate of 14 (85 mol % by GC, 11.8 g, 67.5 mmol)
was added mercaptoacetic acid (8.8 g, 95.5 mmol) in a flask
open to the atmosphere, and the mixture was stirred for 16
h (an initial exotherm occurs raising the internal temperature
of the mixture to approximately 60 °C) giving crude 13 (20.6
g, approximately 73 mol % 13 by HPLC; 19:1 mol ratio of
13:13a). A portion of this product (17.9 g) was suspended
in water (50 mL), and a solution of potassium carbonate (96.8
g, 700 mmol) in water (240 mL) slowly added. To the
resulting stirred mixture was added a solution of potassium
peroxymonosulfate (OXONE, 130 g, 211 mmol) in water
(450 mL) over a 2 h period (the internal temperature rose to
45 °C). The suspension was stirred for 16 h, water (300 mL)
was added to give a solution, and this was washed with
diethyl ether (2 × 200 mL, these washings were discarded)
and acidified by slow addition of 3 M sulfuric acid
(approximately 300 mL). The mixture was then extracted
with diethyl ether (2 × 200 mL), and the combined extracts
were washed with brine (50 mL), dried over anhydrous
sodium sulfate, and concentrated to give a clear oil. After
1
1277, 1107, 711; H NMR (400 MHz, DMSO-d6) δ 8.00-
7.97 (m, 2H), 7.70-7.66 (m, 1H), 7.54 (t, J ) 6.2 Hz, 2H),
7.29-7.25 (m, 2H), 7.22-7.16 (m, 3H), 4.64 (t, J ) 4.6
Hz, 2H), 3.65 (t, J ) 4.6 Hz, 2H), 3.56 (t, J ) 5.6 Hz, 2H),
3.48 (t, J ) 5.0 Hz, 2H), 3.21-3.17 (m, 2H), 2.77 (t, J )
5.6 Hz, 2H), 1.96-1.92 (m, 2H); 13C NMR (100 MHz,
DMSO-d6) δ 165.3, 138.9, 133.5, 129.2, 128.7, 128.1, 125.9,
70.8, 67.8, 58.1, 51.1, 50.4, 35.4, 21.7. Anal. Calcd for
C20H24O5S: C, 63.81; H, 6.43; S, 8.52. Found: C, 63.89;
H, 6.39; S, 8.39.
4-Hydroxy-7-{2-[(2-{[3-(2-phenylethoxy)propyl]-
sulfonyl}ethyl)amino]ethyl}-1,3-benzothiazol-2(3H)-one Hy-
drochloride (1). Ethanol (473 kg), 18 (50.4 kg, 134 mol),
and 2 (30.0 kg, 122 mol) were charged to a suitable reaction
vessel with vigorous stirring. Triethylamine (43.1 kg, 435
mol) was added, and the mixture was refluxed (75 °C or
above) for around 4.5 h. The reaction mixture was cooled
to approximately 65 °C, and 36% hydrochloric acid (55.5
kg, 549 mol) was added at a controlled rate keeping the
temperature of the mixture below 72 °C during the addition.
Residual hydrochloric acid was washed in with ethanol (15
kg), and the temperature of the mixture was adjusted to 65-
69 °C. Seed crystals of 1 (approximately 0.90 kg) were
added, and the stirred mixture was allowed to cool to 10-
640
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

dissolution in toluene (150 mL) and concentration (rotary
evaporator, water bath 50 °C, water pump), the oil was
dissolved in toluene (150 mL) at 60 °C, and the solution
was left to stand for 16 h. The product that had crystallized
was collected by filtration and dried to give 91 (10.3 g, 61%)
as a white solid, HPLC purity 99.3%.
169.2, 151.1, 141.6, 138.6, 136.0, 126.8, 124.6, 108.6, 56.0,
38.7, 34.6, 22.6.
N-[2-(4-Methoxy-2-oxo-2,3-dihydrobenzothiazol-7-yl)-
ethyl]acetamide (20). A stirred suspension of 8 (42.3 g,
0.149 mol) in 36% hydrochloric acid (212 mL) was heated
to reflux giving a red-brown solution. The solution was
refluxed for 2 h 20 min during which time a precipitate had
formed. The mixture was allowed to cool, and the precipi-
tated product was collected by filtration, air-dried, and then
dried further under vacuum for 16 h affording crude 20 (20.3
g, 51%) as a yellow solid, HPLC purity 87%. This material
was dissolved in ethanol (160 mL) and water (160 mL) at
76-80 °C, and the resulting solution was filtered hot and
allowed to cool with stirring for 16 h. The recrystallized
product was collected by filtration, washed with 2-propanol,
and dried under vacuum for 16 h to give 20 (15.1 g, 38%
overall) as a yellow solid, HPLC purity 96%; MS (APCI
+ve) m/z 267 (M+ + H); 1H NMR (400 MHz, DMSO-d6) δ
11.85 (s, 1H), 7.93 (m, 1H), 6.95 (d, J ) 8.4 Hz, 1H), 6.92
(d, J ) 8.4 Hz, 1H), 3.85 (s, 3H), 3.25 (q, J ) 6.8 Hz, 2H),
2.63 (t, J ) 6.8 Hz, 2H), 1.77 (s, 3H); 13C NMR (75 MHz,
DMSO-d6) δ 169.5, 169.1, 143.4, 125.3, 125.2, 123.4, 122.8,
108.8, 55.9, 38.7, 33.5, 22.6.
7-(2-Aminoethyl)-4-methoxy-1,3-benzothiazol-2(3H)-
one, Hydrochloride (21a). A suspension of trifluoroacet-
amide 33 (10 g, 0.029 mol) in a mixture of ethanol (100
mL), water (50 mL), and 36% hydrochloric acid (50 mL)
was heated to reflux and refluxed for 4 h. The resulting
brown solution was concentrated on a rotary evaporator
(water bath at 40 °C), removing most of the ethanol to initiate
crystallization of the product. The resulting mixture was
cooled at 0-4 °C for 35 min, and the product was isolated
by filtration, washed with diethyl ether (around 60 mL), and
dried under vacuum at 50 °C for 16 h to give 21a (5.48 g,
71%) as a fine yellow powder, HPLC purity 95%; MS (FAB
+ve) m/z 225 (M+ + H); 1H NMR (300 MHz, DMSO-d6) δ
8.24 (br s, 3H), 7.03 (d, J ) 8.4 Hz, 1H), 6.96 (d, J ) 8.4
Hz, 1H), 3.86 (s, 3H), 3.08-2.93 (br m, 2H), 2.91-2.82
(m, 2H); 13C NMR (100 MHz, DMSO-d6) δ 179.3, 144.0,
125.8, 123.4, 122.9, 108.9, 56.0, 38.0, 31.1.
Preparation of Vinyl Sulfone 15 from Acid 9. To a
solution of 9 (2.86 g, 10 mmol) in THF (10 mL) was added
1 M borane‚THF solution (14 mL, 14 mmol) resulting in
effervescence which subsided over 10 min. The mixture was
stirred at ambient temperature for 3 h and then a solution of
sodium hydroxide (1.63 g) in water (5 mL) was added
(effervescence). After the mixture was stirred for 30 min,
saturated brine (3 mL) was added, and the mixture was
extracted with diethyl ether (50 mL). The organic phase was
separated, dried over anhydrous sodium sulfate, and con-
centrated to give crude 17 as a clear oil (2.22 g). A portion
of this product (2.0 g) was dissolved in toluene (2 mL),
thionyl chloride (0.9 mL, 12.3 mmol) was added, and the
mixture was heated at 95-100 °C for 1.5 h. After cooling,
the reaction mixture was concentrated to give an oil which
was dissolved in dichloromethane (10 mL) and added
dropwise to a solution of triethylamine (2.1 mL, 15 mmol)
in dichloromethane (1.5 mL). The resulting mixture was
stirred for 2 h, then washed with 2 M hydrochloric acid (20
mL) and brine (10 mL), dried over anhydrous sodium sulfate,
and concentrated to give the crude product as an oil which
was purified by chromatography on silica gel eluting with
10% ethyl acetate/90% dichloromethane. This afforded pure
15 (0.68 g, 30%) as a clear oil which crystallized on standing,
mp 28-28.5 °C, HPLC purity 99.5%; MS (FAB +ve) m/z
255 (M+ + H); 1H NMR (360 MHz, CDCl3) δ 7.30 (t, J )
7.0 Hz, 2H), 7.25-7.14 (m, 3H), 6.57 (dd, J ) 16.6, 9.7
Hz, 1H), 6.42 (d, J ) 16.6 Hz, 1H), 6.14 (d, J ) 9.7 Hz,
1H), 3.64 (t, J ) 7.0 Hz, 2H), 3.52 (t, J ) 5.9 Hz, 2H), 3.02
(m, 2H), 2.87 (t, J ) 7.0 Hz, 2H), 2.01 (m, 2H).
N-[2-(2-Bromo-4-methoxybenzothiazol-7-yl)ethyl]acet-
amide (19). To 48% HBr solution (44 mL) at 22 °C with
stirring was added 8 (12.0 g, 0.042 mol), and the resulting
red-brown solution was heated to 100 °C (internal temper-
ature) over a 55 min period using controlled heating on a
jacketed vessel. The reaction mixture was quenched by
addition of water (500 mL) with stirring giving a precipitate,
and the mixture was cooled in an ice bath and neutralized
by portionwise addition of 28% ammonia solution. The
resulting mixture was stirred in an ice bath for 30 min
resulting in formation of a brown solid that was isolated by
filtration and washed with 2-propanol to give crude 19 as a
brown solid (9 g). This material was dissolved in ethyl acetate
(160 mL) by refluxing for 2 h and filtered hot, and the filtrate
was allowed to cool to ambient temperature with stirring.
The recrystallized product was collected by filtration, washed
with 2-propanol, and dried under vacuum for 16 h giving
19 (4.8 g, 34% overall), HPLC purity 90%; MS (APCI +ve)
m/z 329 and 331 (M+ + H); 1H NMR (300 MHz, CDCl3) δ
7.17 (d, J ) 8.1 Hz, 1H), 6.86 (d, J ) 8.1 Hz, 1H), 5.47 (br
s, 1H), 4.02 (s, 3H), 3.58 (d, J ) 6.9 Hz, 2H), 2.95 (t, J )
6.9 Hz, 2H), 1.95 (s, 3H); 13C NMR (75 MHz, DMSO-d6) δ
N-[2-(4-Hydroxy-2-oxo-2,3-dihydrobenzothiazol-7-yl)-
ethyl]acetamide (22). To a suspension of 2 (20.4 g, 0.082
mol) in DMF (100 mL) was added triethylamine (27.5 mL,
0.196 mol) at 10 °C, and the mixture was stirred for 10 min.
Acetyl chloride (7.0 mL, 0.098 mol) was added dropwise
over 30 min at 10-20 °C, and the mixture was allowed to
warm to ambient temperature and stirred for 3 h. The mixture
was cooled to 10 °C and water (100 mL) was added over a
10 min period resulting in the formation of a precipitate.
The suspended solid was stirred for 30 min, then collected
by filtration, washed with water (200 mL), and dried under
vacuum at 45 °C for 16 h to give crude 22 (15.6 g, 75%) as
a yellow solid. Purification could be achieved by recrystal-
lization from methanol/water (75/25 v/v, 43% recovery for
the first crop) to afford 22 as an off-white solid, HPLC purity
1
96.1%; MS (EI) m/z 252 (M+ + H); H NMR (300 MHz,
DMSO-d6) δ 11.50 (br s, 1H), 9.94 (br s, 1H), 7.92 (m, 1H),
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
641

6.78 (d, J ) 8.4 Hz, 1H), 6.70 (d, J ) 8.4 Hz, 1H), 3.23 (m,
2H), 2.58 (t, J ) 7.2 Hz, 2H), 1.98 (s, 3H).
up and validation of processes to 2), Orgamol SA (scale-up
and validation of processes to 2), and Rhodia-Chirex
((formerly Chirex Ltd.) improved nitration process and scale-
up of processes to 2). Finally, Robert Heald, Magnus
Blomberg, Caroline Nutley, Anil Mistry, and many other
AstraZeneca scientists are acknowledged for their support
of this project.
Acknowledgment
We thank Gerald Steele, Steve Cosgrove, and Ingvar
Yme´n and co-workers for their studies on solid-state
characterization of 1 and their insightful suggestions. Kenneth
Jansson, Tomas Kjellqvist, Karin Tho¨rnblom, Erica Johans-
son, Jacek Bielawski, and Helena Hedqvist deserve thanks
for their contributions to the successful scale-up, transfer,
and validation of the final-stage process to the production
facility and its subsequent validation. The following com-
panies are thanked for their work on the nine outsourced
registered stages: Merck-Lipha SA (improved process for
oxidation of 16, scale-up and validation of processes to 18),
Siegfried CMS AG (scale-up of processes to 18), EMS-
Dottikon AG (improved deprotection process for 5, scale-
Supporting Information Available
Table summarizing critical parameters and activities
associated with the registered processing stages of 1. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Received for review March 2, 2004.
OP049953Y
642
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development