Kepner-Tregoe Decision Analysis as a tool to aid route selection. Part 3. Application to a back-up series of compounds in the PDK project

Article abstract of DOI:10.1021/op8000355

Kepner-Tregoe Decision Analysis was used to rank 22 potential routes to a back-up series of compounds in the PDK project. The ten highest scoring routes were evaluated practically, affording four new synthetic sequences for preparing the target compounds.

Full text of DOI:10.1021/op8000355

Organic Process Research & Development 2008, 12, 1060–1077
Kepner-Tregoe Decision Analysis as a Tool To Aid Route Selection. Part 3.
Application to a Back-Up Series of Compounds in the PDK Project
Jeremy S. Parker,* John F. Bower, Paul M. Murray, Bharti Patel, and Pere Talavera
AstraZeneca, Process Research and DeVelopment, AVlon Works, SeVern Road, Hallen, Bristol BS10 7ZE, U.K.
Abstract:
Kepner-Tregoe Decision Analysis was used to rank 22 potential
routes to a back-up series of compounds in the PDK project. The
ten highest scoring routes were evaluated practically, affording
four new synthetic sequences for preparing the target compounds.
Introduction
Figure 1. PDK back-up compounds.
Scheme 1. Medicinal Chemistry route
The use of Kepner-Tregoe Decision Analysis (KTDA) to
help prioritise synthetic routes has been described in a previous
paper in this series.1 Following on from the lead compound in
the PDK project, AZD7545,2 a back-up series of compounds
was identified for the treatment of type II diabetes (Figure 1).
The back-up series bears some similarities to AZD7545,
the most important being the presence of the same chiral
amide group. The back-up compounds are also anilines
with a p-sulfone and an o-chlorine; however, the sulfones
were changed to simple alkyl and a piperazine unit was
introduced in the meta position.
The medicinal chemistry synthetic route was a seven-step
synthesis starting from 2,3,4-trifluoronitrobenzene (Scheme 1
shows the preparation of AZM574670). The 2-fluorine of the
starting material was selectively displaced using ammonia in a
sealed tube, and then the 4-fluorine was substituted using the
appropriate alkyl thiol. The amino group was converted to a
chloro using a Sandmeyer reaction, and the nitro group was
reduced. The chiral amine portion of the molecule was then
introduced using the original conditions developed for AZD7545.
Finally the sulfide was oxidised to the afford the sulfone, and
the appropriate piperazine was introduced using an SNAr
reaction to afford the target molecules. All four possible
piperazines were commercially available, and sourcing these
molecules on scale was not considered to be an issue.
When the project was received into Process Research
and Development, this route was evaluated for potential
long-term manufacture, allowing identification of the key
issues. The most significant of the two issues identified
was that the third step of the synthesis is a diazotization,
and following a hazard evaluation of this reaction, it was
decided that it would not be safe for further scale-up using
standard processing equipment. The other major issue was
that the starting material was extremely expensive, and
using this compound would push the cost of the final API
well beyond the cost of goods target.
Having determined that a new synthesis would be
required, a search for alternative routes was started. The
preparation of 1,2,3,4-tetrasubstituted aromatic rings is a
challenging synthetic problem, and following extensive
literature searching, 21 new routes were proposed by
scientists across the department.
The use of Kepner-Tregoe Decision Analysis was then
introduced to compare the ideas, allowing prioritisation of routes
and development of a work plan.
Kepner-Tregoe Decision Analysis
Decision Statement. Following discussions by the project
team, the following decision statement was agreed upon: “To
identify routes for the long-term manufacture of the drug
candidates, the bond-forming reactions of which will be used
for delivery of the first campaign, paying particular attention
to safety, health and environmental issues. Any new routes
* To whom correspondence should be addressed. E-mail: jeremy.parker@
astrazeneca.com.
(1) Parker, J. S.; Moseley, J. D. Org. Process Res. DeV. 2008, 12, 1041–
1043.
(2) Moseley, J. D.; Brown, D.; Firkin, C.; Jenkin, S.; Patel, B.; Snape, E.
Org. Process Res. DeV. 2008, 12, 1044–1059.
1060
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development
10.1021/op8000355 CCC: $40.75
2008 American Chemical Society
Published on Web 10/28/2008

Table 2. Wants
should have the flexibility to deliver any of the candidates under
consideration.”
factor
Want
The statement clarifies that the objective is to identify routes
that could be used for long-term manufacture and that whichever
route was selected, the intention would be to use transformations
identified to deliver the first campaign.
accommodation
routes with no difficult reactions (e.g., low
temperature steps, exothermic reactions, etc)
well precedented chemistry
feasibility
number of steps
flexibility
minimum number of steps
to be able to deliver any of the compounds
by introducing the variable portions late in
the synthesis
Develop Objectives. The list of possible factors was
reviewed,1 and the following were selected:
no. of steps to key step
a minimum number of steps to key step
• Safety, health and environment were obvious factors
to include, particularly as they were identified in the
decision statement.
• Intellectual property was included, as any new route
would need to have freedom to operate if outside the
scope of the medicinal chemistry patent.
• Accommodation of any difficult steps needed to be
considered.
• Raw material costs were important, as starting material
costs made the medicinal chemistry route too expensive.
Additionally, any proposed starting materials needed
to be available to purchase in bulk.
preparation of common intermediates while the short list was
being narrowed to one compound.
Weigh the Wants. The relative importance of the Wants
was considered, and feasibility, number of steps and number
of steps to key step were selected as the most important. It was
hoped that these represented a balance between how likely the
route was to work (feasibility) and how quickly the route could
be evaluated (number of steps to key step) against the benefits
of reducing the length of the route (number of steps). Accom-
modation was considered to be the next most important Want,
scoring 7, with flexibility being less important, scoring 4 (Table
3).
• Feasibility of any proposed chemistry needed to be
considered.
• Number of steps, with a reduction from the seven steps
used in the medicinal chemistry route desired.
• Number of steps to key step was included, as with
limited resource it was unlikely that it would be possible
to evaluate all of the routes proposed. Thus, working
through the options as quickly as possible by tackling
key reactions, the maximum number of routes could
be evaluated.
Table 3. Weighting for Wants
Want
weighting
accommodation
feasibility
number of steps
flexibility
7
10
10
4
number of steps to key step
10
• Flexibility was included as the project had yet to choose
a single compound, and thus routes allowing delivery
of all four compounds were required.
Generate Alternatives. A request for ideas was issued
across the Global Process Research and Development depart-
ment, resulting in 21 alternative routes to the target compounds.
With these alternatives in hand the process of scoring, ordering
and selecting new routes for evaluation could begin.
Identify Musts and Wants. The factors selected in the
Develop Objects section were considered and divided into Musts
(Table 1) and Wants (Table 2). It was decided that flexibility
was both a Must and a Want, as it was essential that a route
could deliver any of the three compounds under consideration
but also desirable that the route introduced the variability
(piperazine and/or alkyl sulfone) late in the synthesis, allowing
Screen Alternatives through the Musts. A data table
(Table 4) was generated to analyse the suggested routes (routes
being displayed by name of starting material, then by route
name), and the alternatives were screened against the Musts.3
Routes passing Musts were scored 1; those failing to meet the
objective scored 0. The majority of failures were due to high
raw material costs, including the medicinal chemistry route and
a number of variants using the same starting material. Some
other routes failed for not being flexible enough to deliver all
the candidates under consideration. Multiplying the scores
together gives a final score of 1 (route passes all Musts) or 0
(route fails at least one Must), giving a go/no go decision for
each route to proceed to scoring against the Wants.
Table 1. Musts
factor
Must
safety
health
have no safety issues that would prevent
large-scale manufacture
have no health issues that would prevent
large-scale manufacture
environment
raw materials
not use stoichiometric heavy metals
use raw materials that are available to
purchase on a large scale
Compare Alternatives against the Wants. A scoring
system for the Wants was now developed to allow the routes
to be evaluated (Table 5):
intellectual property have no patent or freedom to operate
issues
flexibility
raw material costs meet raw material cost targets
The cost of goods for the compounds was
allow delivery any of the compounds
• Accommodation. It was decided to consider the final
three steps of any route (those most likely to be
conducted in-house in the future), scoring 10 for any
routes with no perceived problems. Three points would
be deducted for any “problem” steps.
known to be an issue. The exact raw ma-
terial cost target was unknown at the start
of the route selection work, so this factor
remained in the Decision Analysis calcu-
lations, but no routes were eliminated due
to high raw material costs unless the cost
equaled or exceed that for the medicinal
chemistry route.
(3) Details of the highest scoring 10 routes are included in this paper.
Details of lower scoring routes and those not passing the Musts are
provided in Supporting Information.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1061

1062
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Table 5. Scoring system for Wants
of an advanced intermediate which could be used to
deliver any of the possible drug candidates. Of the two
fragments, late introduction of the piperazine section
was seen the more important, as this fragment was the
more complex. Thus, any route that introduced both
fragments in the final three steps scored 10, if the
piperazine was introduced in the final three steps, the
route scored 5 and if neither fragment was introduced
in the final three steps the route scored 0.
• Number of steps to key step. Number of steps to key
step measured how quickly it would be possible to look
at the critical transformations in each route, either
eliminating the route if the transformation did not work
or increasing the feasibility of the route if the trans-
formation did work. Routes where the key step was the
first step (thus could be evaluated quickly) were scored
10, with the score decreasing the further through the
route that the key step lay. With the scoring system
want
scoring system
accommodation
consider final 3 steps
score 10
-3 for any “problem” steps
(sulfide to sulfone oxidation
considered to be a problem step)
score 10 for exact literature
or AstraZeneca precedent
score 8 for close literature
precedent
feasibility
score 6 for a “better than
even chance”
score 4 for significant doubts,
but some precedent
score 2 for “one obscure paper
suggests this is possible”
score 10 for 3 steps
number of steps
score 9 for 4 steps
score 8 for 5 steps
score 7 for 6 steps
completed, all routes were scored against the Wants (Table 5),
and the scores multiplied by the weight of the Want, giving a
total score for each route. The routes were then ranked by total
score, the highest scoring 360, the lowest scoring 189.
Identify Adverse Consequences. The top scoring routes
were assessed in the light of the decision-making process, and
no issues that would adversely affect the investigation of these
routes were identified.
Make the Best Balanced Choice. The decision analysis
procedure produced an ordered list of 14 routes that had the
potential to deliver the drug candidates (Table 4). With limited
amount of time and resource available, it was decided to
investigate the top 10 routes (those scoring 249 or above), and
work started initially focusing on the highest scoring
score 6 for 7 steps
score 5 for 8 steps
score 4 for 9 steps
score 3 for 10 steps
score 2 for 11 steps
score 1 for 12 steps
score 0 for >12 steps
score 10 if both heterocycle and
sulfur introduced in last 3 steps
score 5 if only heterocycle
introduced in last 3 steps
score 0 if neither heterocycle or
sulfur introduced in last 3 steps
flexibility
number of steps to key step score 10 for first step
score 9 for second step
score 8 for third step
score 7 for fourth step
score 6 for fifth step
routes.
score 5 for sixth step
score 4 for seventh step
score 3 for eighth step
score 2 for ninth step
Results and Discussion
The investigations into each of the 10 highest scoring
routes are discussed below. For each proposed route, a
scheme (for the synthesis of AZM574670) is given
covering the type of chemistry required for each trans-
formation. Additionally the Key Step in each scheme is
marked in red. The chemistry that had been demonstrated
previously (medicinal chemistry route) is shown in blue
boxes. A table providing comments on how the route was
evaluated against the Wants is available in the Appendix.
Route A
score 1 for 10th step
score 0 for >10th step
• Feasibility. The feasibility of the transformations was
considered an extremely important factor in comparing
the routes. The highest scoring routes would be any in
which all the transformations had been demonstrated,
and of the 22 routes proposed, only the medicinal
chemistry route fulfilled this criteria. Thus, the other
routes were scored against literature or in-house pre-
cedence (Table 5), with those that appeared to be
reasonably feasible scoring highly and those with more
speculative ideas scoring lower.
• Number of steps. Reducing the number of steps is a
key aim in looking for an alternative route. It was
initially hoped that it would be possible to identify a
three-step route to the target compounds, and a score
of 10 was offered for any route that achieved this. On
receiving the alternative routes, the shortest option was
four steps, which scored 9 (it was decided not adjust
the scoring for the lack of three-step routes), and as
the number of steps increased, the score decreased.
• Flexibility. Flexibility had been selected as a Want as
it was hoped that it would be possible to find a route
where the piperazine and sulfone fragments were
introduced at a late stage, allowing the early preparation
The highest scoring route, Route A (see Table A), is a four-
step route starting from 1-chloro-2,6-difluoro-benzene (5). This
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1063

material is initially sulfonylated ortho to the fluorine (either
fluorine as the molecule is symmetrical). The key step is then
selectively displacing the fluorine para to the alkylsulfone group
with an ammonia source. For this amination reaction, it was
hoped that the bulk of the alkylsulfone group would shield the
o-fluorine, improving the selectivity for substitution of the
p-fluorine. If this reaction was successful, the sequence could
be completed by acylating the aniline and introducing the
piperazine using an SNAr reaction.
tion at the o-fluorine, affording 2-chloro-3-(ethylsulfonyl)-6-
fluoroaniline (8) (Scheme 4), with no evidence of any substi-
tution at the p-fluorine.
Scheme 4
The sulfonylation reaction was investigated using
standard conditions from the literature.4 Thus, 1-chloro-
2,6-difluoro-benzene (5) was reacted with ethanesulfonyl
chloride, catalysed by aluminium trichloride. Initial studies
used nitrobenzene as the solvent, but when no conversion
was observed, the reaction was conducted without solvent
at 100 °C. Slow consumption of the starting material was
observed using these conditions, affording a mixture of
products including small amounts of the required 2-chloro-
1,3-difluoro-4-(ethylsulfonyl)benzene (6), as well as mono-
and bis-chlorination of the starting material.
With such good selectivity for substitution at the
o-fluorine, it was decided to introduce the N-methyl
piperazine group first, and then investigate the conversion
of the p-fluorine to an amino group. However, reacting
2-chloro-1,3-difluoro-4-(ethylsulfonyl)benzene (6) with
N-methyl piperazine in THF gave exclusive substitution
at the p-fluorine affording 1-(2-chloro-4-(ethylsulfonyl)-
3-fluorophenyl)-4-methylpiperazine (9) (Scheme 5), with
no substitution at the o-fluorine.
In an attempt to avoid the chlorination of the starting
material, it was decide to investigate the use of the correspond-
ing fluoride in the reaction.5 Ethanesulfonyl fluoride is not
commercially available, but methanesulfonyl fluoride can be
purchased and was used to investigate this approach. Also
investigated at this time was the use of sulfonic anhydrides
catalysed by trifluoromethanesulfonic acid.6 Ethanesulfonic
anhydride is not commercially available, but methanesulfonic
anhydride could be purchased. Both these reactions were
successful in producing 2-chloro-1,3-difluoro-4-(methylsulfo-
nyl)benzene (7), and the procedure using the sulfonic anhydride
was chosen for further development (Scheme 2) as it was
simpler to handle the reagents and waste products.
Scheme 5
The two reactions showed that it was possible to get
selective reaction at both the o- and p-fluorines, but the
regioselectivity with ammonia and N-methyl piperazine
was opposite to that required. To get around this problem,
it was decided to identify an ammonia surrogate that would
give the required regioselectivity at the p-fluorine and
could then be converted to the required aniline. Work
started looking at primary amines, but benzylamine,
benzhydrylamine and tert-butylamine all reacted exclu-
sively at the o-fluorine, and no reaction was observed with
tritylamine. Attention then changed to looking at second-
ary amines, as N-methyl piperazine (a secondary amine)
had given exclusive reaction at the p-fluorine. No reaction
was observed with dibenzylamine, 1:1 ortho:para selec-
tively was observed with diallylamine and 2:5 ortho:para
selectivity with isoindoline. The selectivity with isoindo-
line was the most promising, and this reaction was probed
further in an attempt to improve the regioselectivity.
Increasing the polarity of the solvent was found to improve
selectivity for para-substitution, with 1:160 ortho:para
selectively in NMP and 1:50 ortho:para selectively in
MeOH. Although good selectivity was observed, it was
decided not to pursue this approach as the removal of the
isoindoline group is known to be challenging,8 and the
Scheme 2
Having identified suitable methodology for the sulfo-
nylation reaction, ethanesulfonic anhydride was prepared7
and reacted with 1-chloro-2,6-difluoro-benzene (5) to
afford 2-chloro-1,3-difluoro-4-(ethylsulfonyl)benzene (6)
(Scheme 3). The yield of the reaction was disappointing
compared to the equivalent reaction with methanesulfonic
anhydride, but enough material was prepared to continue
with the route investigation.
Scheme 3
(4) Truce, W. E.; Vriesen, C. W. J. Am. Chem. Soc. 1953, 75, 5032–
5036.
The key step of the synthesis was then investigated, looking
to substitute the p-fluorine to afford the required aniline.
Reacting 2-chloro-1,3-difluoro-4-(ethylsulfonyl)benzene (6) with
aqueous ammonia in organic solvents gave exclusive substitu-
(5) Hyatt, J. A.; White, A. W. Synthesis 1984, 214–217.
(6) Ono, M.; Nakamura, Y.; Sato, S.; Itoh, I. Chem. Lett. 1988, 395-
398.
(7) Field, L.; Settlage, P. H. J. Am. Chem. Soc. 1954, 76, 1222-
1225.
1064
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

presence of a chlorine atom on the molecule was likely
to further complicate this. Thus attention refocused on the
use of diallylamine. Changing the solvent to NMP and
keeping the temperature low gave 1:40 ortho:para selec-
tively. The allyl groups could then be removed using
palladium (0)-catalyzed deallylation9 to afford the re-
quired 2-chloro-4-(ethylsulfonyl)-3-fluoroaniline (10)
(Scheme 6).
Route B
Scheme 6
Having obtained the required intermediate, it was decided
to reverse the order of the final two steps, introducing the chiral
amine portion as the final step. Thus, 2-chloro-4-(ethylsulfonyl)-
3-fluoroaniline (10) was reacted with N-methyl piperazine,
affording 2-chloro-3-(4-methylpiperazin-1-yl)-4-(ethylsulfonyl)-
aniline (11) (Scheme 7).
The route with the second highest score, Route B (see Table
B), was one step longer than Route A but scored highly due to
the key step being the first in the sequence, as well having
precedented chemistry. Starting with 2,3-dichlorotoluene (14),
the sulfone group is introduced using electrophilic aromatic
substitution chemistry. The methyl group is then oxidised, and
the piperazine is introduced by nucleophilic aromatic substitu-
tion of chlorine.10 Finally, a Hoffman reaction converts the
carboxylic acid group to an amine, which can be acylated to
afford the required product.
Scheme 7
Scheme 9
The final step, introduction of the chiral amide portion
of the molecule, was achieved using the original conditions
developed for the preparation of AZD7545.2 (R)-3,3,3-
Trifluoro-2-hydroxy-2-methyl-propionic acid (12) was
doubly protected with trimethylsilyl groups and then
converted to (S)-3,3,3-trifluoro-2-methyl-2-trimethylsila-
nyloxy-propionyl chloride (13) using oxalyl chloride. This
acid chloride was then reacted with 2-chloro-3-(4-meth-
ylpiperazin-1-yl)-4-(ethylsulfonyl)aniline (11) (Scheme 8),
affording AZM574670 (1) and completing the first new
route to the drug candidates.
Workstarted looking at the key electrophilic aromatic
substitution reaction, chemistry that was also under investigation
for Route A. In Route B, regioselectivity was necessary to set
up the correct substitution pattern on the aromatic ring, and thus
sulfonylation para to the methyl group was required. It was
hoped that the ortho/para directing effect of both the methyl
group and the 3-chloride would provide some selectivity at the
4-position and that it would be possible to purify the product
from any regioisomeric byproduct. The reaction was initially
investigated using ethane sulfonyl chloride catalysed by alu-
minium trichloride in benzotrifluoride. The procedure produced
a complicated mixture of products, and it was decided to change
to using the sulfonylation conditions successfully demonstrated
in Route A. Reacting 2,3-dichlorotoluene (14) with methane-
sulfonic anhydride in the presence of trifluoromethanesulfonic
acid cleanly produced three products in a ratio of 2:1:2. Analysis
of these products identified them as the 4-, 5- and 6-sulfones
(Scheme 9) with substitution dominating at the 4- and 6-posi-
tions.
Scheme 8
With no selectivity for 4-sulfonylation observed in the key
reaction, it was decided to halt work on Route B.
(9) Garro-Helion, F.; Merzouk, A.; Guibe´, F. J. Org. Chem. 1993, 58,
6109–6113.
(8) Hou, D-R.; Hsieh, Y-D; Hsieh, Y-W. Tetrahedron Lett. 2005, 46,
5927–5929.
(10) Feit, P. W.; Tvaermose Nielsen, O. B. J. Med. Chem. 1976, 19, 402–406.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1065

Scheme 12
Route C
Two steps remained to link Route C with the medicinal
chemistry route, the first of which was reduction of the nitro-
group to an aniline. Hydrogenation was considered, but with
the aromatic halogens present, it was decided to investigate
alternative methodologies. Sodium hydrosulfite was successfully
used as a reducing agent,12 affording 2-chloro-4-(ethylsulfonyl)-
3-fluoroaniline (10) (Scheme 13).
Scheme 13
Route C (see also Table C) represented the closest route to
the original medicinal chemistry route, with the aniline being
replaced with the equivalent phenol. The phenol had the
advantage of being less than 25% of the cost of aniline and
removed the hazardous diazotisation step from the route.
The first step on the route was completed by reacting the
sodium salt of ethanethiol with 2,3-difluoro-6-nitrophenol (15),
affording 3-(ethylthio)-2-fluoro-6-nitrophenol (16) in a 90%
yield (Scheme 10).
Finally, this compound was acylated using the conditions
developed for AZD7545,2 affording (2R)-N-[2-chloro-4-(eth-
ylsulfonyl)-3-fluorophenyl]-3,3,3-trifluoro-2-hydroxy-2-meth-
ylpropanamide (19) (Scheme 14) and thus linking into the
medicinal chemistry route.
Scheme 14
Scheme 10
Completion of this synthetic sequence represented a second
alternative route to the drug candidates.
Route D
Chlorination of this molecule was investigated using phos-
phorus oxychloride and a variety of bases;11 however, no
reaction was observed. A further review of the literature
suggested that the presence of electron-withdrawing groups on
the aromatic ring is important to allow the reaction to proceed.
Thus, it was decided to reverse the order of steps 2 and 3, first
oxidising the sulfide to the sulfone to reduce the electron density
of the aromatic ring, and then investigate the chlorination.
3-(Ethylthio)-2-fluoro-6-nitrophenol (16) was oxidised to 3-(eth-
ylsulfonyl)-2-fluoro-6-nitrophenol (17), using the conditions
developed for AZD75452 (Scheme 11).
Scheme 11
The chlorination was investigated again using phosphorus
oxychloride and a variety of bases, which identified triethy-
lamine as the best base for the reaction, affording 2-chloro-4-
(ethanesulfonyl)-3-fluoro-1-nitrobenzene (18) in an 81% yield
(Scheme 12).
The key step in Route D (see Table D) is metalation of the
aromatic ring, followed by chlorination using an electrophilic
chloride source, such as NCS. Fluorine has been shown to be
a powerful ortho director of lithiation,13 and it was anticipated
that the fluorine in the 3-position would direct lithiation to the
2-position.
(11) Sbarskii, V. L.; Shutov, G. M.; Zhilin, V. F.; Chirkova, R. G.; Orlova,
E. Y. J. Org. Chem. USSR. (Eng. Transl.) 1971, 7, 303–307.
(12) Antolini, L. J. Heterocycl. Chem. 1992, 29, 1449–1455.
1066
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

The first step on Route D was based on the chemistry reported
by Suggs,14 displacing the 1-fluorine of 1,2-difluoro-4-nitrobenzene
(20) using the sodium salt of ethanethiol and affording 1-(ethylthio)-
2-fluoro-4-nitrobenzene (21) (Scheme 15). The major impurity
observed in the reaction was the formation of the bisthioether (22).
With the difficulties separating product from remaining starting
material, it was decided to investigate alterative substrates for the
metalation/chlorination reaction. The intention was to find a
substrate that would give complete chlorination, avoiding the
problems of separation of chlorinated and nonchlorinated com-
pounds. It was reasoned that introduction of the N-methyl pipera-
zine group at the 2-position may provide some chelation to the
lithium base, providing more effective lithiation of the aromatic.
Thus, it was decided to investigate the displacement of the
remaining fluorine on 1-(ethylthio)-2-fluoro-4-nitrobenzene (21).
This approach would introduce the N-methyl piperazine group
earlier than planned but, if successfully metallated/chlorinated,
would afford a molecule with the correct heteroatoms in the 1-4
positions at an early stage. Displacement of the fluorine was found
to proceed using conditions similar to those used on the medicinal
chemistry route, stirring the starting material in neat N-methyl
piperazine for 24 h at 140 °C (Scheme 17). The majority of the
starting material was consumed, but significant amounts of impuri-
ties were generated in the reaction, including aniline (25) (4%)
and azo compound (26) (20%). However, the required product
could be easily separated using flash column chromatography
affording 1-[2-(ethylthio)-5-nitrophenyl]-4-methylpiperazine (24)
in a 55% yield.
Scheme 15
The metalation/chlorination of both 1,2-difluoro-4-nitrobenzene
(20) and 1-(ethylthio)-2-fluoro-4-nitrobenzene (21) was then in-
vestigated using s-BuLi and NCS. The reaction afforded a mixture
of compounds, including a significant amount of starting material,
and analysis suggested that metalation was taking place at various
locations on the aromatic ring, as well as addition of butyl groups.
It was thus decided to investigate the metalation/chlorination
reaction stepwise, first looking at the extent and location of the
lithiation and then examining the chlorination. The lithiation
reaction was investigated using LDA and LHMDS, followed by
quenching with acetic acid-d, with the incorporation of deuterium
(determined by 1H and 2D NMR) measuring the extent and location
of deprotonation. LDA proved to be too strongly basic, causing
decomposition of both 1,2-difluoro-4-nitrobenzene (20) and 1-(eth-
ylthio)-2-fluoro-4-nitrobenzene (21). LHMDS was effective at
lithiating 1-(ethylthio)-2-fluoro-4-nitrobenzene (21), with the major-
ity of deuteration being observed at the 3-position (54% deutera-
tion); however, deuteration was also observed at the 5-position
(22% deuteration) and to a lesser extent at the 6-position (9%
deuteration). The most effective combination of base and substrate
was LHMDS and 1,2-difluoro-4-nitrobenzene (20), where deu-
teration was observed exclusively at the 3-position (50% deutera-
tion), with no decomposition of the starting material. The use of
these conditions, then quenching the lithiated aromatic with NCS,
afforded a 60-70% conversion to 2-chloro-3,4-difluoro-1-ni-
trobenzene (23) (Scheme 16). Purification of the product proved
to be challenging, with separation by flash column chromatography
difficult due to the similar polarity of the remaining starting material
and product. However, with care, 20-30% of the required product
could be isolated for use in further transformations.
Scheme 17
As before, the metalation/chlorination of 1-[2-(ethylthio)-5-
nitrophenyl]-4-methylpiperazine (24) was investigated in a step-
wise manner, using s-BuLi, LDA and LHMDS followed by a
deuterium quench to probe the extent and location of lithiation.
With s-BuLi, extensive decomposition of the starting material was
observed, forming a mixture of unknown products. No deuterium
incorporation was observed using LDA or LHMDS, suggesting
that substrate was not undergoing any lithiation using these bases.
The limited success in investigating the metalation/chlorina-
tion approach focused the work onto the successful conversion
of 1,2-difluoro-4-nitrobenzene (20) to 2-chloro-3,4-difluoro-1-
nitrobenzene (23), and it was decided to progress the product
of this reaction through the required transformations to link into
the medicinal chemistry route. Thus, the displacement of the
4-fluorine of 2-chloro-3,4-difluoro-1-nitrobenzene (23) using so-
dium ethanethiolate was investigated. Conversion of the starting
material to 2-chloro-4-(ethylthio)-3-fluoro-1-nitrobenzene (27) was
observed (44%, Scheme 18), but the final reaction mixture also
contained thioether (28) (14%), formed by displacement of the
2-chlorine rather than the 4-fluorine and bisthioether (29) (16%),
as well as residual starting material (23) (7%).
Scheme 16
(13) Gilday, J. P.; Negri, J. T.; Widdowson, D. A. Tetrahedron 1989, 45,
4605–4618.
(14) Farmer, J. D.; Rudnicki, S. M.; Suggs, J. W. Tetrahedron Lett. 1988,
29, 5105–5108.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1067

Scheme 18
Scheme 20
As with the substrates on Route D, it was decide to initially
investigate the extent and location of lithiation, and thus
1-(ethylthio)-2,4-difluorobenzene (32) was reacted with LDA
and LHMDS, followed by quenching with acetic acid-d. LDA
was found to deprotonate exclusively at the 3-position (30%
deuteration), with no decomposition of the starting material.
No deprotonation was observed using LHMDS.
As the lithiation of 1-(ethylthio)-2,4-difluorobenzene (32)
was not as extensive as that of 1,2-difluoro-4-nitrobenzene (20)
evaluated for Route D (30% vs 50%), it was decided to pursue
the lithiation/chlorination of the later substrate, thus halting work
on Route E.
As with previous compounds on this route, purification of
the product by flash column chromatography was challenging
due to the similar polarity of the impurities, and only 26% of
the required 2-chloro-4-(ethylthio)-3-fluoro-1-nitrobenzene (27)
was isolated.
Completing work on the route, the nitro-group was reduced
to an amino-group using hydrogenation. A catalyst was sourced
from Johnson-Matthey for the reduction of aromatic nitro-
groups in the presence of aromatic chlorides, and this success-
fully converted 2-chloro-4-(ethylthio)-3-fluoro-1-nitrobenzene
(27) to 2-chloro-4-(ethylthio)-3-fluoroaniline (30) (Scheme 19).
Route F
Scheme 19
This final conversion linked Route D to the medicinal chemistry
route, formally providing a third route to the drug candidates.
Route E
Route F was another four-step route, with two SNAr reactions
introducing the alkylsulfone and the heterocycle. The key step
in the sequence was the second SNAr reaction (fourth step), as
this involved the displacement of a chlorine atom. The
equivalent reaction, displacing a fluorine atom, had been
demonstrated as part of the medicinal chemistry route, but the
lower reactivity of the chlorine16 made it essential to investigate
this transformation (see also Table F).
The first step of the route was investigated using sodium
ethanesulfinate17 and 1,2,3-trichloro-4-nitrobenzene (33)
based on conditions from the literature.18 No significant
reaction was observed in ethanol and THF, but slow
consumption of the starting material was noted in DMSO
and DMF at 130 °C. Analysis of the reaction in DMSO
after 24 h showed that the majority of the starting material
remained (86%), with a small amount of the required 2,3-
dichloro-1-(ethylsulfonyl)-4-nitrobenzene (34) present (2%,
Scheme 21). The analysis also detected small amounts of
1,2-dichloro-3-(ethylsulfonyl)-4-nitrobenzene (35, 2%),
2-dichloro-1,3-bis(ethylsulfonyl)-4-nitrobenzene (36, 2%)
and other sulfonylated species.
Route E has a key step similar to that of Route D but starts
from 2,4-difluorothiophenol (31), affording a substrate with two
ortho directing fluorines.13 The key metalation/chlorination step
was examined at the same time as those for Route D (see also
Table E).
2,4-Difluorothiophenol (31) was alkylated with bromoet-
hane,15 affording 1-(ethylthio)-2,4-difluorobenzene (32) in a
79% yield (Scheme 20).
On reflection, this result is not particularly surprising, as
introduction of the first ethylsulfonyl group further activates
the ring towards nucleophilic aromatic substitution, promoting
(15) Wnuk, S. F.; Robins, M. J. J. Org. Chem. 1990, 55, 4757–4760.
(16) Smith, M. B.; March, J. AdVanced Organic Chemistry; Wiley
Interscience: New York, 2001; Chapter 13.
(17) Bearlocher,; F, J.; Baerlocher, M. O.; Chaulk, C. L.; Langler, R. F.;
MacQuarrie, S. L. Aust. J. Chem. 2000, 53, 399–402.
1068
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Scheme 21
Scheme 22
Route H
the introduction of further ethylsulfonyl groups. With the lack
of selectivity and poor conversion for the first step, it was
decided to stop work on Route F.
Route G
Route H pursues a strategy similar to that of Route D, with
the key step involving the metalation/chlorination of the
aromatic ring. The route is made one step shorter than Route
D by introduction of the sulfur substituent at the correct
oxidation state. However, the route was viewed as having poorer
feasibility, as the acidic protons on the alkyl group of the sulfone
were likely to interfere with the metalation/chlorination step
(see also Table H).
The first step of the route was investigated at the same time
as the equivalent first step on Route F, using sodium ethane-
sulfinate to displace the halogen para to the nitro-group. As
with route F, no significant reaction was observed in ethanol
and THF, and slow consumption of the starting material was
noted in DMSO and DMF at 130 °C. Further investigation
of the reaction in DMSO, increasing the amount of the
sodium ethanesulfinate (1 equiv to 5 equiv), resulted in
the consumption of the majority of the starting 1,2-
difluoro-4-nitrobenzene (38), affording the required 1-(eth-
ylsulfonyl)-2-fluoro-4-nitrobenzene (39), but with a larger
amounts of the bisulphone (40)(Scheme 23).
Route G shows similarities to Route F, both having four
steps and with the same final two transformations. Route G
differs from Route F with aniline nitrogen starting at the correct
oxidation state and the sulfur being oxidised in the second step,
whereas Route F has the sulfur at the correct oxidation state
with the nitrogen being reduced in the second step. As with
Route F, there were concerns about the displacements of the
chlorine, and thus the fourth step was considered to be the key
step (see also Table G). A similar issue to that encountered in
Route B was also present, with regioselectivity being required
in the first-step alkylthiation.
The alkylthiation chemistry is known in the literature, but
only moderate selectivity is seen between the positions ortho
and para to the aniline nitrogen.19 With the presence of the
two additional chlorines, it was hoped that better selectivity
would be observed para to the amino-group. The alkylthiation
reaction of 2,3-dichloroaniline (37) was investigated using
standard conditions and resulted in a number of new products.
Analysis of these showed that alkylthiation had occurred
extensively at the 4- and 6-positions (Scheme 22), as well some
dialkylthiation.
Scheme 23
With the lack of selectivity for the 4-thiol and the anticipated
difficulty in separating the regioisomers, it was decided to stop
work on the route.
Analogous to the first reaction on Route F, the
introduction of the first ethylsulfonyl group further
activates the ring towards nucleophilic aromatic substitu-
tion, promoting the introduction of the second ethylsul-
fonyl group. As fluorine is a better leaving group than
chlorine, the reaction proceeds to give significant amounts
of both the mono- and bisulfones, whereas in Route F,
only small amounts of sulfonylated products are observed.
(18) (a) Li, J. J.; Norton, M. B.; Reinhard, E. J.; Anderson, G. D.; Gregory,
S. A.; Isakson, P. C.; Koboldt, C. M.; Masferrer, J. L.; Perkins, W. E.;
Seibert, K.; Zhang, Y.; Zweifel, B. S.; Reitz, D. B. J. Med. Chem.
1996, 39, 1846–1856. (b) Dack, K. N.; Dickinson, R. P.; Long, C. J.;
Steele, J. Bioorg. Med. Chem. Lett. 1998, 8, 2061–2066. (c) Lacour,
J.; Monchaud, D.; Bernardinelli, G.; Favarger, F. Org. Lett. 2001, 3,
1407–1410.
(19) Ranken, P. F.; McKinnie, B. G. J. Org. Chem. 1989, 54, 2985–2988.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1069

The lack of selectivity observed in the first step of Route
H halted work on this approach.
Route I
G, the key step in the sequence was seen to the be the
displacement of chloride by the nitrogen heterocycle, and
as the last transformation in the six-step sequence, Route
J scored poorly on number of steps to key step (see Table
J).
Studies on the route started with the thiocyanation reaction
of 2,3-dichloroaniline (37). These reactions are well precedented
to give substitution para to the nitrogen of anilines, and using
the AZD7545 conditions, the correct regioisomer was obtained
in a 76% yield (Scheme 24), with no trace of any ortho
substitution.
Scheme 24
The obtained 4-amino-2,3-dichlorophenyl thiocyanate (41)
was then reduced and alkylated in a one-pot procedure using
sodium sulfide and bromoethane, affording 2,3-dichloro-4-
(ethylthio)aniline (42) (Scheme 25).
Route I bears a number of similarities to Route D, with
the key step being the metalation of the aromatic ring,
followed by halogenation. Work on Routes D and E had
been completed at the time of starting work on Route I,
and the experience gained on the key transformation of
these two reactions was considered. With the difficultly
of regioselective metalation of the aromatic substrate in
Routes D and E and also without any ortho-directing
fluorides present, it was felt unlikely that it would be
possible to get regioselective bromination in the key step
of Route I. Thus the feasibility score for the route was
changed from 4 to 2, reducing the total score and ranking
of Route I. For this reason it was decided not start work
on Route I (see also Table I), transferring the available
effort onto Route J.
Scheme 25
Oxidation of the sulfur was achieved using hydrogen
peroxide/sodium tungstate, affording 2,3-dichloro-4-(ethylsul-
fonyl)aniline (43) (Scheme 26).
Scheme 26
Route J
The acylation to introduce the chiral amide portion was
completed using the AZD7545 conditions,2 affording (2R)-N-
[2,3-dichloro-4-(ethylsulfonyl)phenyl]-3,3,3-trifluoro-2-hydroxy-
2-methylpropanamide (44) (Scheme 27).
Scheme 27
The final sequence evaluated as part of the route
selection was Route J. The final three steps in this route
were identical to those in Route G, but three steps were
required to introduce the thioethyl substrate, whereas only
one step was required in Route G. It was felt that the
thiocyanation, reduction and alkylation sequence was well
precedented, as the same chemistry had been used for the
preparation of AZD7545.2 However, as with Routes F and
Finally, the key nucleophilic aromatic substitution of the
3-chlorine was investigated, using the medicinal chemistry
conditions for the displacement of the equivalent fluorine. The
reaction was successful, affording AZM574670 (1) in a 78%
yield (Scheme 28).
The completion of this sequence represented the fourth new
route to the drug candidates and marked the end of the
experimental work on the project.
1070
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Scheme 28
Scheme 29
Data Review. The completion of the synthetic work
provided an opportunity to reassess all the routes against the
Wants that had been selected, showing how the proven routes
measured up against each other (Table 6).
Scheme 30
All routes now score 10 for feasibility, as all the steps had
been demonstrated. Additionally Route A had changed from 4
to 5 steps, decreasing the number of steps score from 9 to 8.
A revaluation of the routes against the Decision Statement
clearly showed that Route D was not suitable for long-term
manufacture as the yields and selectivity of the first two
reactions are poor, as well as the first step requiring low
temperatures. The other routes all had potential, and the relative
merits of each needed to be considered.
A further review of the Wants used to prioritise the
evaluation of routes showed that on completion of synthetic
work, both feasibility and steps to key step were no longer
relevant (the former scoring 10 for all routes and the latter not
being applicable as all the key steps had been proven).
Accommodation and flexibility did not provide any differentia-
tion between the routes, as the same score was obtained for
each route. This left number of steps, which is an important
factor for choosing between proven routes, as it implicitly
contains elements of throughput, yield and cost. Thus a simple
method of choosing an appropriate route for manufacture
involved measuring the number of steps, which gave Route A
as the preferred route, as it has the smallest number of steps.
An alternative approach (which was not conducted due to
termination of the project) would be to conduct a new Kepner-
Tregoe Decision Analysis session to compare the proven routes,
using a set of criteria based on the fact that detailed information
was available for all the transformations. Factors such as overall
yield and robustness, as well as number of steps, could be used
to provide a more considered comparison of the proven routes.
Scheme 31
ment, Scheme 31) and Route J (2,3-dichloroaniline/thiocyana-
tion, Scheme 32).
Conclusions
A further evaluation of the Wants used to prioritise the routes
suggests that Route A would be preferred route for manufacture.
A more through comparison of the proven routes could be
completed using Kepner-Tregoe Decision Analysis, incorporat-
ing additional factors that had been measured during the
experimental evaluation.
The investigation into the 10 highest scoring routes afforded
four new alternatives to the drug candidates: Route A (1-chloro-
2,6-difluorobenzene/sulfonylation route, Scheme 29), Route C
(2,3-difluoro-6-nitrophenol/phenol activation and elimination,
Scheme 30), Route D (3,4-difluoronitrobenzene/thiol displace-
Table 6. Final Decision Analysis data table
Wants
route title
accommodation, 7a
feasibility, 10a
no. of steps, 10a
flexibility, 4a
steps to key step, 10a
total score
Route A
Route C
Route D
Route J
10
10
10
10
10
10
10
10
8
7
7
7
5
5
5
5
9
8
9
5
360
340
350
310
a Weighting.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1071

Scheme 32
difluoro-4-(methylsulfonyl)benzene (7) as a brown solid
(11.08 g, 73%). GC purity 59%; tR 12.1 min; δH (400
MHz, d6-DMSO) 7.89 (m, 1H), 7.57 (m 1H), 3.38 (s, 3H);
δC (100 MHz, d6-DMSO) 161.30 (dd, J ) 256.3, 3.1 Hz),
155.54 (dd, J ) 257.0, 3.8 Hz), 128.87 (d, J ) 10.7 Hz),
126.05 (dd, J ) 14.6, 3.8 Hz), 113.19 (dd, J ) 21.5, 3.8
Hz), 110.55 (t, J ) 21.5 Hz), 43.68 (d, J ) 2.3 Hz); δF
(376 MHz, d6-DMSO) -103.24 (m), -108.21 (m); MS
(EI+) 226/228 (M+).
Preparation of 2-Chloro-1,3-difluoro-4-(ethylsulfonyl)ben-
zene (6). Ethanesulfonic anhydride (1.36 g, 6.73 mmol, 2.00
equiv) was charged to a small reaction tube with trifluo-
romethanesulfonic acid (0.60 mL, 6.73 mmol, 2.00 equiv),
and the mixture was heated to 70 °C. 1-Chloro-2,6-
difluorobenzene (5) (0.50 g, 3.37 mmol, 1.00 equiv, tR
3.1 min) was added in a single portion, and the mixture
then placed in an oil bath at 130 °C for 18 h. The reaction
was then cooled to 55 °C, and water (10 mL) was added
(note: addition is exothermic), followed by ethyl acetate
(10 mL). The layers were separated, and the aqueous layer
was washed with ethyl acetate (10 mL). The organic
extracts were combined and washed with aqueous sodium
hydrogen carbonate solution (saturated, 10 mL), water (10
mL) and aqueous sodium chloride solution (saturated, 10
mL), and the solvent was removed in vacuo, affording a
dark brown oil which was purified by flash column
chromatography, affording 2-chloro-1,3-difluoro-4-(eth-
ylsulfonyl)benzene (6) as a pale brown solid (109 mg,
13%). GC purity 100%; tR 13.0 min; δH (400 MHz, CDCl3)
7.89 (m, 1H), 7.21 (m, 1H), 3.34 (q, J ) 7.4 Hz, 2H),
1.33 (t, J ) 7.4 Hz, 3H); δC (100 MHz, CDCl3) 162.39
(dd, J ) 260.3, 3.1 Hz), 156.51 (dd, J ) 258.7, 3.8 Hz),
129.39 (d, J ) 10.0 Hz), 123.80 (d, J ) 16.1 Hz), 123.86
(d, J ) 4.4 Hz), 112.49 (dd, J ) 22.3, 3.8 Hz), 112.01
(d, J ) 21.9 Hz), 50.16 (d, J ) 2.3 Hz), 6.98; δF (376
MHz, CDCl3) -101.79 (m), -106.99 (m); MS (CI+) 226/
228 (MH+).
Preparation of 2-Chloro-4-(ethylsulfonyl)-3-fluoroaniline
(10). 2-Chloro-1,3-difluoro-4-(ethylsulfonyl)benzene (6)
(366.0 mg, 1.52 mmol, 1.00 equiv) was charged to a small
reaction tube with diallylamine (0.75 mL, 6.08 mmol, 4.00
equiv) and N- methylpyrrolidone (3.6 mL). The tube
contents were stirred at ambient temperature for 7 days.
The reaction mixture was partitioned between N-butyl
acetate (20 mL) and water (20 mL). The layers were
separated, and the organic portion was washed with water
(20 mL). The organic solution was dried over magnesium
sulfate and filtered, and the solvent was removed in vacuo,
affording a yellow oil. The oil was redissolved in
dichloromethane (2 mL), and this solution was added to
a small reaction tube containing palladium(II) acetate (6.95
mg, 0.31 mmol, 0.02 equiv), triphenylphosphine (32.5 mg,
0.124 mmol, 0.08 equiv) and dichloromethane (2 mL)
under a nitrogen atmosphere. N,N-Dimethylbarbituric acid
(1.21 g, 7.74 mmol, 5.00 equiv) was added in a single
portion, and the mixture was stirred at 45 °C for 4 days.
The solvent was then removed in vacuo, and the orange
oil partitioned between ethyl acetate (20 mL) and saturated
Disappointingly at this stage in the exercise, the project was
terminated because of adverse toxicology results for all of the
compounds, and the selection of a single route for long-term
manufacture was never completed. However, this has provided
an opportunity to publish the work conducted in this area using
Kepner-Tregoe Decision Analysis, providing an extensive
example of how such a technique can be used to aid route
selection.
Experimental Section
General Procedures. Melting points were determined using
a Griffin melting point apparatus (aluminium heating block)
1
and are uncorrected. H, 13C and 19F NMR spectra were
recorded on a Varian Inova 400 MHz spectrometer with
chemical shifts given in ppm, using internal standards (1H and
13C) and external standards (19F). The reaction mixtures and
products were analysed by gas chromatography on a Hewlett-
Packard 6890 according to the following conditions: injector,
250 °C; split ratio, 25:1; column, Agilent HP-5, 5% phenyl
methyl siloxane, Capillary 30 m × 320 µm × 1.00 mm
nominal; pressure, 80 kPa, make-up, nitrogen; temperature
program, 50-0-10-280-7; detector, 250 °C. GC purities are
quoted as normalized area %. Analytical TLC was carried out
on commercially prepared plates coated with 0.25 mm of self-
indicating Merck Kieselgel 60 F254 and visualised by UV light
at 254 nm. Preparative scale silica gel flash chromatography
was carried out by standard procedures using Merck Kieselgel
60 (230-400 mesh).
Route A. Preparation of 2-Chloro-1,3-difluoro-4-(methyl-
sulfonyl)benzene (7). Methanesulfonic anhydride (23.46 g,
134.7 mmol, 2.00 equiv) was charged to a 100 mL flask
with trifluoromethanesulfonic acid (0.596 mL, 6.73 mmol,
0.10 equiv), and the mixture heated to 70 °C. 1-Chloro-
2,6-difluorobenzene (5) (10.00 g, 67.4 mmol, 1.00 equiv,
tR 3.1 min) was added over 2 min, and the mixture then
placed in an oil bath at 130 °C for 21 h. The reaction was
then cooled to 55 °C, and water (50 mL) was added (note:
addition is exothermic), followed by ethyl acetate (50 mL).
The layers were separated, and the aqueous layer was
washed with ethyl acetate (25 mL). The organic extracts
were combined and washed with aqueous sodium hydro-
gen carbonate solution (saturated, 2 × 50 mL) and aqueous
sodium chloride solution (saturated, 25 mL), and the
solvent was removed in vacuo, affording 2-chloro-1,3-
1072
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

aqueous sodium hydrogen carbonate solution (10 mL). The
layers were separated, the organic portion was dried over
magnesium sulfate and filtered, and the solvent was
removed in vacuo, affording an orange oil which was
purified by flash column chromatography (eluent 50%
EtOAc in isohexane), affording 2-chloro-4-(ethylsulfonyl)-
3-fluoroaniline (10) as pale yellow solid (198 mg, 55%).
GC purity 99%; tR 18.8 min; mp 113.5-114.5 °C; υmax
cm-1 3456, 3634, 1301, 1132; δH (400 MHz, d6-DMSO)
1.08 (t, J ) 7.3 Hz, 3H), 3.22 (q, J ) 7.3 Hz, 2H), 6.67
(d, J ) 9.0 Hz, 1H), 6.72 (s, 2H), 7.37 (dd, J ) 8.3, 9.0
Hz, 1H); δC (100 MHz, d6-DMSO) 7.12, 49.84, 103.72
(d, J ) 19.9 Hz), 109.29 (d, J ) 2.3 Hz), 111.67 (d, J )
14.6 Hz), 128.89 (d, J ) 1.5 Hz), 151.90 (d, J ) 3.1 Hz),
155.98 (d, J ) 250.9 Hz); MS (EI+) 238/240 (M+).
Preparation of 2-Chloro-3-(4-methylpiperazin-1-yl)-4-(ethyl-
sulfonyl)aniline (11). 2-Chloro-4-(ethylsulfonyl)-3-fluoroa-
niline (10) (50.0 mg, 0.210 mmol, 1.0 equiv) was charged
to a small reaction tube with 1-methylpiperazine (1.00 mL,
9.00 mmol, 42.8 equiv). The tube was sealed, and the
mixture was stirred at 130 °C for 16 h. The solvent was
then removed in vacuo, and the brown oil was partitioned
between ethyl acetate (5 mL) and saturated aqueous
sodium hydrogen carbonate solution (5 mL). The layers
were separated, the organic portion was dried over
magnesium sulfate and filtered, and the solvent was
removed in vacuo, affording 2-chloro-3-(4-methylpiper-
azin-1-yl)-4-(ethylsulfonyl)aniline (11) as a yellow solid
(55 mg, 82%). GC purity 100%; tR 24.6 min; δH (400
MHz, CDCl3) 7.77 (d, J ) 8.5 Hz, 1H), 6.70 (d, J ) 8.7
Hz, 1H), 4.58 (br s, 2H), 3.83 (m, 2H), 3.47 (q, J ) 7.5
Hz, 2H), 2.85 (m, 4H), 2.36 (s, 3H), 2.31 (m, 2H), 1.24
(t, J ) 7.6 Hz, 3H); δC (100 MHz, CDCl3) 149.27, 148.21,
130.18, 127.15, 120.22, 111.75, 55.32, 49.00, 48.62, 46.46,
7.47; MS (ES+) 318/320 (MH+).
was added, and the mixture was stirred for a further 30 min.
The layers were separated, and the organic portion further
extracted with aqueous hydrochloric acid (2 M, 5 mL). The
combined aqueous extracts were basified with aqueous sodium
hydroxide solution (2 M, 20 mL) and extracted with ethyl
acetate (2 × 20 mL). The combined organic extracted were
dried over magnesium sulfate and filtered, and the solvent was
removed in vacuo, affording (2R)-N-[2-chloro-4-(ethylsulfonyl)-
3-(4-methylpiperazin-1-yl)phenyl]-3,3,3-trifluoro-2-hydroxy-2-
methylpropanamide (1) as a cream solid (101 mg, 92%) GC
purity 95%; tR 28.7 min; δH (400 MHz, CDCl3) 9.90 (s, 1H),
8.61 (d, J ) 9.0 Hz, 1H), 8.01 (d, J ) 9.0 Hz, 1H), 4.04 (m,
1H), 3.89 (m, 1H), 3.46 (q, J ) 7.3 Hz, 2H), 2.97 (m, 4H),
2.48 (m, 5H), 1.75 (s, 3H), 1.28 (t, J ) 7.4 Hz, 3H); δC (100
MHz, CDCl3) 167.23, 146.69, 140.05, 134.08, 130.39, 125.41,
124.30 (q, J ) 286.4 Hz), 118.06, 75.75 (q, J ) 28.4 Hz), 55.17,
49.09, 47.89, 47.57, 46.05, 20.38, 7.37; MS (ES+) 458/460
(MH+).
Route C. Preparation of 3-(Ethylthio)-2-fluoro-6-nitrophenol
(16). To a suspension of sodium ethanethiolate (5.00 g, 2.83
equiv, 47.6 mmol) in tetrahydrofuran (25 mL) at 0 °C, under a
nitrogen atmosphere was added a solution of 2,3-difluoro-6-
nitrophenol (15) (3.00 g, 1.00 equiv, 16.8 mmol) in tetrahy-
drofuran (25 mL) over 5 min to give a brown slurry. The
mixture was stirred at room temperature overnight. The mixture
was then partitioned between ethyl acetate (50 mL), aqueous
hydrochloric acid (1 M, 50 mL) and aqueous sodium chloride
solution (saturated, 50 mL). The layers were separated, and the
organic layer was washed with aqueous sodium chloride
solution (saturated, 2 × 15 mL), dried over magnesium sulfate,
and filtered, and the solvent was removed in vacuo, affording
a brown solid, which was purified by flash column chroma-
tography (eluent 10% EtOAc in isohexane) affording 3-(eth-
ylthio)-2-fluoro-6-nitrophenol (16) as an orange oil (2.05 g,
90%). GC purity 90%; tR 14.5 min; mp (crude) 70.5-71.5 °C;
Preparation of (2R)-N-[2-Chloro-4-(ethylsulfonyl)-3-(4-me-
thylpiperazin-1-yl)phenyl]-3,3,3-trifluoro-2-hydroxy-2-methyl-
propanamide (1). (R)-3,3,3-Trifluoro-2-hydroxy-2-methyl-
propionic acid (12) (876 mg, 5.54 mmol) was dissolved in
dichloromethane (12.75 mL) and dimethylformamide (21.3 µL).
1,3-Bis(trimethylsilyl)urea (1.42 g, 6.93 mmol) was added
portionwise, followed by dichloromethane (12.75 mL), and the
mixture was stirred at ambient temperature for 24 h. The mixture
was then filtered, and the collected solid was washed with
dichloromethane (2 × 5.8 mL). Oxalyl chloride (696.24 mg,
5.49 mmol) in dichloromethane (2 mL) was added to the filtrate
over 1 h, and the mixture was stirred at ambient temperature
for 24 h. Finally, the reaction solution was filtered to remove
the remaining urea byproducts, affording (S)-3,3,3-trifluoro-2-
methyl-2-trimethylsilanyloxy-propionyl chloride (13) as a solu-
tion in dichloromethane. 2-Chloro-3-(4-methylpiperazin-1-yl)-
4-(ethylsulfonyl)aniline (11) (73.6 mg, 0.24 mmol) was dissolved
in dichloromethane (1 mL), and a portion of the previously
prepared (S)-3,3,3-trifluoro-2-methyl-2-trimethylsilanyloxy-pro-
pionyl chloride (13) solution in dichloromethane (2.4 mL) was
added in a single portion, followed by triethylamine (69.7 mg,
0.69 mmol). The mixture was stirred at ambient temperature
for 90 min, and then aqueous hydrochloric acid (2 M, 5 mL)
υ
max cm-1 3231, 1512, 1204, 1143; δH (400 MHz, d6-DMSO)
1.29 (t, J ) 7.3 Hz, 3H), 3.11 (q, J ) 7.3 Hz, 2H), 7.00 (dd, J
) 6.9, 9.2 Hz, 1H), 7.79 (dd, J ) 1.8, 9.2 Hz, 1H), 11.14 (s,
1H); δC (100 MHz, d6-DMSO) 13.78, 24.42, 115.50, 120.54
(d, J ) 3.1 Hz), 133.68 (d, J ) 14.6 Hz), 134.76 (d, J ) 3.1
Hz), 141.26 (d, J ) 16.9 Hz), 148.24 (d, J ) 239.4 Hz); MS
(EI+) 217 (M+).
Preparation of 3-(Ethylsulfonyl)-2-fluoro-6-nitrophenol (17).
To a solution of 1-(ethylthio)-2-fluoro-4-nitrobenzene (16) (2.10
g, 1.00 equiv, 9.66 mmol) in acetic acid (20 mL) was added
aqueous hydrogen peroxide solution (27.5%, 3.00 mL, 2.74
equiv, 26.9 mmol) in acetic acid (10 mL). The mixture was
heated to 75 °C, causing a slight darkening of the solution. After
stirring for 3 h, the reaction was allowed to cool to ambient
temperature. Sodium metabisulfite (1.50 g, 7.70 mmol) in water
(20 mL) was added dropwise to quench excess peroxide/peracid,
the absence of which was confirmed by the use of peroxide
test strips. The solution was partitioned between ethyl acetate
(200 mL) and water (250 mL). The layers were separated, the
organic layer was washed with water (2 × 100 mL), dried over
magnesium sulfate, and filtered, and the solvent was removed
in vacuo, affording 3-(ethylsulfonyl)-2-fluoro-6-nitrophenol (17)
as a yellow solid (2.23 g, 93%). GC purity 97%; tR 16.6 min;
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1073

mp 110.5-111.5 °C; υmax cm-1 3279, 3066, 1538, 1320, 1221,
1129; δH (400 MHz, d6-DMSO) 1.14 (t, J ) 7.4 Hz, 3H), 3.43
(q, J ) 7.4 Hz, 2H), 7.34 (dd, J ) 6.3, 8.8 Hz, 1H), 7.88 (dd,
J ) 1.7, 8.8 Hz, 1H); δC (100 MHz, d6-DMSO) 6.72, 49.43,
117.92, 120.32 (d, J ) 3.8 Hz), 129.99 (d, J ) 13.0 Hz), 141.56
(d, J ) 17.6 Hz), 142.63, 149.94 (d, J ) 253.2 Hz); MS (EI+)
248 (M+).
at ambient temperature overnight. The reaction was quenched
by dropwise addition of water (20 mL) and then further diluted
with ethyl acetate (20 mL). The layers were separated, the
organic layer was washed with water (10 mL) and aqueous
sodium hydrogen carbonate solution (saturated, 2 × 10 mL),
dried over magnesium sulfate, and filtered, and the solvent was
removed in vacuo, affording (2R)-N-[2-chloro-4-(ethylsulfonyl)-
3-fluorophenyl]-3,3,3-trifluoro-2-hydroxy-2-methylpropana-
mide (19) as a pale yellow solid (111 mg, 70%). GC purity
96%; tR 20.6 min; mp 146.5-148 °C; υmax cm-1 3456, 3316,
1692, 1155, 1100; δH (400 MHz, d6-DMSO) 1.12 (t, J ) 7.4
Hz, 3H), 1.60 (s, 3H), 3.40 (q, J ) 7.4 Hz, 2H), 7.83 (dd, J )
8.3, 9.0 Hz, 1H), 8.08 (s, 1H), 8.14 (d, J ) 9.0 Hz, 1H), 9.99
(s, 1H); δC (100 MHz, d6-DMSO) 6.86, 19.41, 49.61, 74.99
(q, J ) 27.9 Hz), 113.49 (d, J ) 21.7 Hz), 117.49 (d, J ) 3.8
Hz), 122.77 (d, J ) 20.0 Hz), 129.20, 140.42, 154.83 (d, J )
254.5 Hz), 167.49; MS (EI+) 376/378 (M+).
Route D. Preparation of 2-Chloro-3,4-difluoro-1-nitroben-
zene (23). 1,2-Difluoro-4-nitrobenzene (20) (12.00 g, 75.4
mmol, 1.00 equiv, tR 5.4 min) was dissolved in tetrahydrofuran
(100 mL) and cooled in a dry ice/acetone bath to -78 °C
(internal reaction temperature -74 °C). Lithium hexamethyl-
disilazide (1.0 M in THF, 75.4 mL, 74.5 mmol, 1.00 equiv)
was added dropwise over 80 min, turning the colourless solution
dark orange. Stirring was then continued for a further 21/2 h.
N-Chlorosuccinamide (10.07 g, 75.4 mmol, 1.00 equiv) in
tetrahydrofuran (200 mL) was added dropwise over 100 min,
turning the reaction mixture first dark burgundy, then back to
dark orange. The mixture was then allowed to warm to ambient
temperature overnight, forming a light brown, cloudy solution.
The reaction solvent was removed in vacuo, affording a brown
solid, which was partitioned between ethyl acetate (400 mL)
and aqueous hydrochloric acid (1 M, 400 mL). The layers were
separated, and the organic layer was washed with aqueous
sodium hydrogen carbonate solution (saturated, 400 mL), water
(400 mL) and aqueous sodium chloride solution (saturated, 400
mL), affording a clear, orange solution. The solvent was
removed from this solution in vacuo, affording a brown oil
which was purified by flash column chromatography (250 g
silica, eluent 1% EtOAc in isohexane) affording 2-chloro-3,4-
difluoro-1-nitrobenzene (23) as a yellow oil (3.96 g, 27%). GC
purity 87%; tR 7.9 min; δH (400 MHz, CDCl3) 7.28 (m, 1H),
7.82 (ddd, J ) 9.4, 4.7, 2.2 Hz, 1H); δC (100 MHz, CDCl3)
113.94 (dd, J ) 22.0, 2.0 Hz), 115.43 (d, J ) 19.2 Hz), 121.23,
144.19, 147.86 (dd, J ) 253.0, 14.0 Hz), 147.86 (dd, J ) 260.0,
13.0 Hz); δF (376 MHz, CDCl3) -125.27 (ddd, J ) 20.4, 7.9,
4.7 Hz), -131.31 (m); MS (EI+) 193 (M+).
Preparation of 2-Chloro-4-(ethanesulfonyl)-3-fluoro-1-ni-
trobenzene (18). 3-(Ethylsulfonyl)-2-fluoro-6-nitrophenol (17)
(500 mg, 1.00 equiv, 2.00 mmol) was charged to a N2-purged
vessel fitted with a water scrubber. Phosphorus oxychloride
(10.0 mL, 52.5 equiv, 105.0 mmol) was then added to form a
yellow solution. Triethylamine (0.85 mL, 3.00 equiv, 6.01
mmol) was added dropwise, affording an orange suspension.
The reaction mixture was then heated to 105 °C and stirred for
60 h. The reaction mixture was then diluted by addition of
toluene (25 mL) and added dropwise into ice-cooled water (150
mL) to quench excess phosphorus oxychloride. Toluene (175
mL) was added, the layers were separated, the organic layer
was washed with water (2 × 100 mL) and aqueous sodium
hydroxide solution (2 M, 100 mL), dried over magnesium
sulfate, and filtered, and the solvent was removed in vacuo,
affording 2-chloro-4-(ethanesulfonyl)-3-fluoro-1-nitrobenzene
(18) as a colourless solid (431 mg, 81%). GC purity 83%; tR
16.9 min; mp 106.5-108.0 °C; υmax cm-1 3096, 1537, 1316,
1143; δH (400 MHz, d6-DMSO) 1.19 (t, J ) 7.3 Hz, 3H), 3.53
(q, J ) 7.3 Hz, 2H), 8.05 (dd, J ) 6.8, 8.6 Hz, 1H), 8.20 (dd,
J ) 1.0, 8.6 ) Hz, 1H); δC (100 MHz, d6-DMSO) 6.71, 49.57,
116.21 (d, J ) 22.9 Hz), 121.43 (d, J ) 4.6 Hz), 129.84, 130.47
(d, J ) 16.2 Hz), 152.81 (d, J ) 254.5 Hz), 156.66; MS (EI+)
267/269 (M+).
Preparation of 2-Chloro-4-(ethylsulfonyl)-3-fluoroaniline
(10). To a suspension of 2-chloro-4-(ethanesulfonyl)-3-fluoro-
1-nitrobenzene (18) (600 mg, 1.00 equiv, 2.25 mmol) in water
(15 mL) was added sodium hydrosulfite (1.40 g, 3.02 equiv,
6.80 mmol). The reaction mixture was then heated to 75 °C
and stirred for 90 min. Aqueous hydrochloric acid (32%, 2.9
mL) was added dropwise to the mixture, and precipitation of
the product was allowed to occur over 3 h by stirring the mixture
at ambient temperature; the mixture was then cooled to 0 °C
for 1 h to promote further precipitation. The reaction mixture
was adjusted to pH 7 by addition of aqueous sodium hydrogen
carbonate (saturated, 25 mL), and the product was isolated by
filtration, washed with water (4 × 10 mL), and dried in a stream
of air affording 2-chloro-4-(ethylsulfonyl)-3-fluoroaniline (10)
as colourless crystals (348 mg, 65%).
Preparation of (2R)-N-[2-Chloro-4-(ethylsulfonyl)-3-fluo-
rophenyl]-3,3,3-trifluoro-2-hydroxy-2-methylpropanamide (19).
To a solution of 2-chloro-4-(ethylsulfonyl)-3-fluoroaniline (10)
(100 mg, 1.00 equiv, 0.42 mmol) in acetonitrile (1.5 mL) at 45
°C was added triethylamine (32 µL, 1.17 equiv, 0.49 mmol) in
acetonitrile (0.2 mL). Thionyl chloride (63 µL, 2.02 equiv, 0.85
mmol) in acetonitrile (0.2 mL) was then added dropwise causing
a darkening of the reaction solution. After 2 h, heating was
stopped, and the vessel was allowed to cool to ambient
temperature. A solution of (2R)-3,3,3-trifluoro-2-hydroxy-2-
methylpropanoic acid (80 mg, 1.21 equiv, 0.51 mmol) in
acetonitrile (0.3 mL) was added, and the reaction was stirred
Preparation of 1-(Ethylthio)-2-fluoro-4-nitrobenzene (21).
1,2-Difluoro-4-nitrobenzene (20) (10.00 g, 62.8 mmol, 1.00
equiv) was dissolved in tetrahydrofuran (100 mL), and the
mixture was sparged with nitrogen for 10 min. Sodium
ethanethiolate (80% purity, 6.61 g, 62.8 mmol, 1.00 equiv) was
added portionwise (Note: addition is exothermic), following
which the mixture was sparged with nitrogen for 10 min. The
mixture was then heated at reflux for 1 h, affording a dark brown
solution. The reaction mixture was then cooled, diluted with
ethyl acetate (150 mL), and washed with aqueous sodium
hydroxide solution (1 M, 2 × 100 mL), aqueous hydrochloric
1074
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

acid solution (1 M, 100 mL), aqueous sodium hydrogen
carbonate solution (saturated, 100 mL) and aqueous sodium
chloride solution (saturated, 100 mL). The solvent was then
removed from the dark brown solution in vacuo, affording a
brown solid which was purified by flash column chromatog-
raphy (250 g silica, eluent 5% EtOAc in isohexane) affording
1-(ethylthio)-2-fluoro-4-nitrobenzene (21) as an orange oil
(10.12 g, 80%). GC purity 90%; tR 13.0 min; δH (400 MHz,
CDCl3) 1.41 (t, J ) 7.4 Hz, 3H), 3.06 (q, J ) 7.3 Hz, 2H),
7.35 (t, J ) 7.9 Hz, 1H), 7.89 (dd, J ) 9.6, 2.2 Hz, 1H), 8.00
(m, 1H); δC (100 MHz, CDCl3) 16.46, 25.54 (d, J ) 2.3 Hz),
110.67 (d, J ) 26.9 Hz), 117.90 (d, J ) 19.0 Hz), 119.59 (d,
J ) 3.1 Hz), 126.87 (d, J ) 3.1 Hz), 135.44 (d, J ) 16.9 Hz),
158.52 (d, J ) 248.6 Hz); δF (376 MHz, CDCl3) -108.45 (m);
MS (EI+) 201 (M+).
Preparation of 1-[2-(Ethylthio)-5-nitrophenyl]-4-methylpip-
erazine (24). 1-(Ethylthio)-2-fluoro-4-nitrobenzene (20) (10.00
g, 49.7 mmol, 1.00 equiv) was dissolved in 1-methyl piperazine
(50.0 mL, 0.45mol, 9.06 equiv), and the mixture was heated
slowly to reflux. The mixture was stirred at reflux for 24 h,
cooled to ambient temperature, diluted with ethyl acetate (250
mL), and washed with water (3 × 250 mL). The solvent was
then removed from the dark brown solution in vacuo, affording
a dark brown solid which was purified by flash column
chromatography (250 g silica, eluent 2% MeOH in CH2Cl2),
affording 1-[2-(ethylthio)-5-nitrophenyl]-4-methylpiperazine
(24) as dark brown oil (7.30 g, 52%). GC purity 100%; tR 21.4
min; δH (400 MHz, CDCl3) 1.43 (t, J ) 7.4 Hz, 3H), 2.38 (s,
3H), 2.62 (br s, 4H), 2.99 (q, J ) 7.4 Hz, 2H), 3.05 (br s, 4H),
7.20 (d, J ) 8.7 Hz, 1H), 7.86 (d, J ) 2.3 Hz, 1H), 7.92 (dd,
J ) 8.7, 2.6 Hz, 1H); δC (100 MHz, CDCl3) 13.22, 25.15, 46.05,
51.27, 55.13, 114.40, 119.12, 123.86, 144.38, 145.15, 149.60;
MS (EI+) 281 (M+).
Preparation of 2-Chloro-4-(ethylthio)-3-fluoro-1-nitroben-
zene (27). 2-Chloro-3,4-difluoro-1-nitrobenzene (23) (2.00 g,
10.3 mmol, 1.00 equiv) was dissolved in tetrahydrofuran (100
mL), and the mixture was sparged with nitrogen for 10 min.
Sodium ethanethiolate (80% purity, 1.09 g, 10.3 mmol, 1.00
equiv) was added in a single portion, following which the
mixture was sparged with nitrogen for 10 min. The mixture
was then heated at reflux for 1 h, affording a dark brown
solution. The reaction mixture was then cooled, and the solvent
was removed in vacuo, affording a dark brown oil. This material
was dissolved in ethyl acetate (100 mL) and washed with
aqueous sodium hydroxide solution (1 M, 100 mL), aqueous
hydrochloric acid solution (1 M, 100 mL), aqueous sodium
hydrogen carbonate solution (saturated, 100 mL) and aqueous
sodium chloride solution (saturated, 100 mL). The solvent was
then removed from the dark brown solution in vacuo, affording
a brown oil, which was purified by flash column chromatog-
raphy (100 g silica, eluent 2% EtOAc in isohexane) affording
2-chloro-4-(ethylthio)-3-fluoro-1-nitrobenzene (27) as an orange
oil (639 mg, 26%). GC purity 73%; tR 15.0 min; δH (400 MHz,
CDCl3) 1.41 (t, J ) 7.1 Hz, 3H), 3.05 (q, J ) 7.4 Hz, 2H),
7.24 (m, 1H), 7.77 (dd, J ) 8.7, 1.5 Hz, 1H); δC (100 MHz,
CDCl3) 13.64, 25.84 (d, J ) 2.3 Hz), 120.10 (d, J ) 3.8 Hz),
121.04 (d, J ) 3.8 Hz), 124.44 (d, J ) 3.1 Hz), 125.32 (d, J )
3.8 Hz), 134.21, 155.23 (d, J ) 248.8 Hz); δF (376 MHz,
CDCl3) -107.28 (d, J ) 6.3 Hz); MS (EI+) 193/195 (M+).
Preparation of 2-Chloro-4-(ethylthio)-3-fluoroaniline (30).
2-Chloro-4-(ethylthio)-3-fluoro-1-nitrobenzene (27) (55.0 mg,
0.23 mmol, 1.00 equiv) was dissolved in ethyl acetate (1.5 mL)
and acetic acid (0.1 mL). Platinum on carbon (JM Pt 117, 5%
w/w, 16.0 mg) was added, and the mixture was stirred at 50
°C under 1barG hydrogen for 18 h. The reaction mixture was
then cooled and filtered to remove the catalyst. The solvent was
then removed from the orange solution in vacuo, affording
2-chloro-4-(ethylthio)-3-fluoroaniline (30) as a red oil (39.0 mg,
81%). GC purity 99%; tR 14.1 min; δH (400 MHz, CDCl3) 1.41
(t, J ) 7.3 Hz, 3H), 3.05 (q, J ) 7.4 Hz, 2H), 4.24 (br s, 2H),
7.24 (m, 1H), 7.77 (dd, J ) 8.7, 1.5 Hz, 1H); δC (100 MHz,
CDCl3) 13.67, 25.88 (d, J ) 2.3 Hz), 111.39 (d, J ) 3.0 Hz),
121.03 (d, J ) 3.8 Hz), 124.48 (d, J ) 3.1 Hz), 133.93 (d, J )
2.0 Hz), 134.19 (d, J ) 18.0 Hz), 155.32 (d, J ) 248.0 Hz); δF
(376 MHz, CDCl3) -107.16 (d, J ) 7.9 Hz); MS (ES+) 206
(M+).
Route E. Preparation of 1-(Ethylthio)-2,4-difluorobenzene
(32). 2,4-Difluorobenzenethiol (31) (1.00 g, 6.84 mmol, 1.00
equiv, tR 4.0 min) was dissolved in ethanol (5 mL) under a
nitrogen atmosphere. Potassium hydroxide (0.426 g, 7.52 mmol,
1.10 equiv) added, and after stirring for 5 min the solid had
completely dissolved. The mixture was stirred for 20 min, then
ethyl bromide (0.77 mL, 10.3 mmol, 1.50 equiv) was added in
a single portion, and the mixture was stirred at ambient
temperature for a further 3.5 h. Aqueous sodium hydroxide
solution (2 M, 4 mL) was then added to destroy any remaining
ethyl bromide, and the ethanol removed in vacuo. The remaining
aqueous solution was extracted with dichloromethane (2 × 10
mL), and the combined extracts washed with water (10 mL),
dried over magnesium sulfate, and filtered, and the solvent was
removed in vacuo, affording 1-(ethylthio)-2,4-difluorobenzene
(31) as a pale yellow oil (0.90 g, 79%). GC purity 98%; tR 6.5
min; δH (400 MHz, CDCl3) 1.25 (t, J ) 7.3 Hz, 3H), 2.86 (q,
J ) 7.3 Hz, 2H), 6.84 (m, 2H), 7.39 (dd, J ) 15.0, 8.1 Hz,
1H); δC (100 MHz, CDCl3) 14.46, 28.36 (d, J ) 1.5 Hz), 104.30
(t, J ) 26.6 Hz), 111.62 (dd, J ) 3.3, 20.1 Hz), 118.14 (d, J )
3.8, 18.4 Hz), 134.51 (d, J ) 3.1, 9.2 Hz), 162.91 (ddd, J )
10.7, 25.7, 249.2 Hz); δF (376 MHz, CDCl3) -104.70 (m),
-111.33 (m); MS (EI+) 174 (M+).
Route J. Preparation of 4-Amino-2,3-dichlorophenyl Thio-
cyanate (41). 2,3-Dichloroaniline (37) (20.00 g, 123.4 mmol,
1.00 equiv) and sodium thiocyanate (20.01 g, 246.9 mmol, 2.00
equiv) were slurried in acetonitrile (200 mL). The mixture was
cooled to 0 °C and bromine (19.73 g, 123.4 mmol, 1.00 equiv)
in acetonitrile (60 mL) was added over 1.5 h. A solution of
sodium hydrogen carbonate (10.37 g, 123.4 mmol, 1.00 equiv)
in water (400 mL) was then added over 30 min. The precipitated
product was stirred at ambient temperature and then isolated
by filtration. The filter cake was washed with water (2 × 200
mL) and then dried in vacuum oven at 40 °C, affording
4-amino-2,3-dichlorophenyl thiocyanate (41) as a white solid
(22.00 g, 76%). GC purity 81%; tR 17.4 min; δH (400 MHz,
d6-DMSO) 7.51 (d, J ) 8.7 Hz, 1H), 6.84 (d, J ) 8.7 Hz, 1H),
6.40 (s, 2H); δC (100 MHz, d6-DMSO) 149.13, 135.63, 134.36,
116.07, 113.88, 111.54, 106.15.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1075

Preparation of 2,3-Dichloro-4-(ethylthio)aniline (42). So-
dium sulfide (13.01 g, 166.6 mmol, 1.80 equiv) was slurried in
dimethylformamide (60 mL), and the mixture was heated to
60 °C. 4-Amino-2,3-dichlorophenyl thiocyanate (41) (15.00 g,
92.6 mmol, 1.00 equiv) in dimethylformamide (52.5 mL) was
added over 10 min, turning the green/blue slurry orange. The
mixture was stirred at 60 °C for 30 min, then bromoethane
(10.59 g, 97.2 mmol, 1.05 equiv) in dimethylformamide (15.0
mL) was added over 10 min, and the mixture was stirred for a
further 3 h. Water (375 mL) was added to the reaction, and the
resultant aqueous solution extracted with methyl tert-butyl ether
(2 × 600 mL). The organic extracts were combined and washed
with water (2 × 600 mL), and the solvent was removed in
vacuo, affording 2,3-dichloro-4-(ethylthio)aniline (42) as a
yellow oil (12.15 g, 61%). GC purity 97%; tR 16.3 min; δH
(400 MHz, CDCl3) 7.18 (d, J ) 8.5 Hz, 1H), 6.63 (d, J ) 8.5
Hz, 1H), 4.16 (br s, 2H), 2.84 (q, J ) 7.3 Hz, 2H), 1.24 (t, J )
7.3 Hz, 3H); δC (100 MHz, CDCl3) 143.54, 135.37, 131.60,
123.42, 118.36, 113.48, 28.68, 14.20; MS (EI+) 222/224/226
(M+).
Preparation of 2,3-Dichloro-4-(ethylsulfonyl)aniline (43).
2,3-Dichloro-4-(ethylthio)aniline (42) (11.00 g, 49.5 mmol, 1.00
equiv) was dissolved in acetonitrile (95 mL). Sodium tungstate
dihydrate (1.47 g, 4.46 mmol, 0.10 equiv) was added, followed
by sulfuric acid (1.46 g, 14.9 mmol, 0.30 equiv) and hydrogen
peroxide solution (25% in water, 20.2 mL, 148.6 mmol, 3.0
equiv). The mixture was stirred at ambient temperature for 18 h,
at which time water (220 mL) was added, causing the
precipitation of a brown solid. The solid was collected by
filtration, washed with water (2 × 50 mL), and dried in a
vacuum oven at 50 °C, affording 2,3-dichloro-4-(ethylsulfony-
l)aniline (43) as a brown solid (12.60 g, 71%). GC purity 97%;
tR 20.5 min; δH (400 MHz, CDCl3) 7.78 (d, J ) 8.7 Hz, 1H),
6.75 (d, J ) 8.7 Hz, 1H), 4.85 (br s, 2H), 3.38 (q, J ) 7.4 Hz,
2H), 1.25 (t, J ) 7.4 Hz, 3H); δC (100 MHz, CDCl3) 148.85,
132.17, 130.74, 124.72, 118.82, 111.83, 48.80, 7.31; MS (EI+)
253/255/257 (M+).
) 29.8 Hz), 48.72, 20.30, 7.12; δF (376 MHz, CDCl3) -80.11;
MS (CI+) 394/396/398 (MH+).
Preparation of (2R)-N-[2-Chloro-4-(ethylsulfonyl)-3-(4-me-
thylpiperazin-1-yl)phenyl]-3,3,3-trifluoro-2-hydroxy-2-methyl-
propanamide (1). (2R)-N-[2,3-Dichloro-4-(ethylsulfonyl)phenyl]-
3,3,3-trifluoro-2-hydroxy-2-methylpropanamide (44) (100 mg,
0.25 mmol, 1.00 equiv) was dissolved in N-methyl piperazine
(0.25 g, 2.50 mmol, 10.0 equiv), and the mixture was heated at
140 °C for 5 h. The reaction mixture was cooled, diluted with
dichloromethane (20 mL), and washed with water (4 × 50 mL).
Isohexane (100 mL) was added to the organic solution, and
the resultant solid collected by filtration and dried in a vacuum
oven at 40 °C, affording (2R)-N-[2-chloro-4-(ethylsulfonyl)-3-
(4-methylpiperazin-1-yl)phenyl]-3,3,3-trifluoro-2-hydroxy-2-
methylpropanamide (1) as a cream solid (112 mg, 78%).
Acknowledgment
Analytical support for all the work described in this paper
was provided by Simon Linke. The authors thank Suzanne
Aitken, Jeff Richards and Nathalie Roberts for help with NMR
analysis and Steve Baldwin for help with MS analysis. The
authors also thank Mike Butters and David Brown for their input
into the project.
Appendix
Each table included below provides comments on how each
route as shown in the text was evaluated against the Wants.
Table A
Preparation of (2R)-N-[2,3-Dichloro-4-(ethylsulfonyl)phe-
nyl]-3,3,3-trifluoro-2-hydroxy-2-methylpropanamide (44). 2,3-
Dichloro-4-(ethylsulfonyl)aniline (43) (8.50 g, 33.4 mmol, 1.00
equiv) was dissolved in acetonitrile (42.5 mL). Triethylamine
(4.90 mL, 35.1 mmol, 1.10 equiv) was added in a single portion,
followed by thionyl chloride (2.56 mL, 35.1 mmol, 1.10 equiv).
(2R)-3,3,3-Trifluoro-2-hydroxy-2-methylpropanoic acid (6.08 g,
38.5 mmol, 1.15 equiv) in acetonitrile (17.0 mL) was added
dropwise, and the mixture was stirred at ambient temperature
for 18 h. The reaction mixture was then partitioned between
aqueous hydrochloric acid solution (0.5M, 170 mL) and
dichloromethane (170 mL). The organic layer was separated,
and the solvent was removed in vacuo, affording a dark brown
oil. This material was triturated with isohexane (200 mL),
affording a brown solid, which was isolated by filtration, washed
with isohexane (2 × 50 mL) and dried in a vacuum oven at 50
°C, affording (2R)-N-[2,3-dichloro-4-(ethylsulfonyl)phenyl]-
3,3,3-trifluoro-2-hydroxy-2-methylpropanamide (44) as a brown
solid (10.50 g, 80%). δH (400 MHz, CDCl3) 9.53 (br s, 1H),
8.61 (d, J ) 9.0 Hz, 1H), 8.06 (d, J ) 8.7 Hz, 1H), 4.09 (s,
1H), 3.43 (q, J ) 7.4 Hz, 2H), 1.78 (s, 3H), 1.28 (t, J ) 7.4
Hz, 3H); δC (100 MHz, CDCl3) 166.56, 139.56, 132.27, 131.66,
130.86, 124.36, 123.63 (q, J ) 287.0 Hz), 118.03, 76.11 (q, J
Table B
Table C
1076
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Table D
Table E
Table F
Table G
Table H
Table I
Table J
Supporting Information Available
Details of lower scoring routes and those not passing the
Musts. This material is available free of charge via the Internet
at http://pubs.acs.org.
Received for review February 15, 2008.
OP8000355
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1077