Process development and pilot-plant synthesis of (2-Chlorophenyl)[2- (phenylsulfonyl)pyridin-3-yl]methanone

Article abstract of DOI:10.1021/op100157q

Routes to (2-chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]methanone, 1, an intermediate in the manufacture of NK1-II inhibitor LY686017 are described which produce 1 in >75% yield and 95% purity. A highly selective telescoped ortho lithation/condensation/oxidation process was developed and successfully scaled to the clinical pilot plant to produce 25 kg of 1. For the pilot-plant campaign, the lithiation step was developed to operate at -50 °C using commercial lithium diisopropylamide (LDA), and the oxidation step employed catalytic TEMPO as the primary and NaOCl as the terminal oxidant. After completion of the pilot-plant campaign second-generation approaches to 1 were developed to improve process greenness where the lithiation and condensation step were operated as warm as -10 °C, the highly efficient AZADO catalyst was used as a substitute for TEMPO in the Anelli-Montanari oxidation, and process mass intensity was reduced 25%.

Full text of DOI:10.1021/op100157q

Organic Process Research & Development 2010, 14, 1229–1238
Process Development and Pilot-Plant Synthesis of
(2-Chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]methanone
Michael E. Kopach,* Michael E. Kobierski, D. Scott Coffey, Charles A. Alt, Tony Zhang, Alfio Borghese, William G. Trankle,
and Dilwyn J. Roberts
Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana 46285, U.S.A.
Humphrey Moynihan and Kurt Lorenz
Eli Lilly SA, Dunderrow, Kinsale, County Cork, Ireland
Orla A. McNamara, Marie Kissane, and Anita R. Maguire
Department of Chemistry and School of Pharmacy, Analytical and Biological Chemistry Research Facility, UniVersity College
Cork, County Cork, Ireland
Scheme 1. NK1-II inhibitor LY686017
Abstract:
Routes to (2-chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]metha-
none, 1, an intermediate in the manufacture of NK1-II inhibitor
LY686017 are described which produce 1 in >75% yield and 95%
purity. A highly selective telescoped ortho lithation/condensation/
oxidation process was developed and successfully scaled to the
clinical pilot plant to produce 25 kg of 1. For the pilot-plant
campaign, the lithiation step was developed to operate at -50 °C
using commercial lithium diisopropylamide (LDA), and the oxida-
tion step employed catalytic TEMPO as the primary and NaOCl
as the terminal oxidant. After completion of the pilot-plant
campaign second-generation approaches to 1 were developed to
improve process greenness where the lithiation and condensation
step were operated as warm as -10 °C, the highly efficient
AZADO catalyst was used as a substitute for TEMPO in the
Anelli-Montanari oxidation, and process mass intensity was
reduced 25%.
methanone, 1, is an important intermediate in the preparation
of LY686017, a potent NK1-II inhibitor which has been studied
clinically for treatment of depression, anxiety, and alcohol
dependency at Eli Lilly and Company (Scheme 1). The synthetic
route for the preparation of 1 has been disclosed and includes
a highly selective ortho lithiation of 2-benzenesulfonyl pyridine,
2, at cryogenic temperatures and then condensation with
2-chlorobenzaldehyde, 4, to produce intermediate alcohol 5.2
The crude alcohol is then efficiently oxidized using the
Anelli-Montanari protocol with catalytic TEMPO as the
primary and NaOCl as the terminal oxidant (Scheme 2). This
process was run in the Lilly clinical pilot plant to produce 25
kg of 1 in five lots with 85% average yield and greater than
95% average purity. Some further optimization was performed
after the pilot-plant campaign was completed to afford an
improved crystallization process which delivered 1 in greater
than 98% purity. Significantly, the process mass intensity3 for
the new process was reduced by 25% relative to the pilot-plant
process. Among the available oxidation methodologies, the
Anelli-Montanari reaction has wide utility and has been used
commercially for oxidation of alcohols to carbonyl compounds.4
The rapid conversions under mild operating conditions are
particularly attractive as an important, emerging green oxidation
methodology. Despite these successes, TEMPO is an expensive
chemical, and for many processes up to 10 mol % is needed to
Introduction
Mammalian tachykinins are peptide neurotransmitters that
interact with central nervous system neurokinin receptors, and
these interactions are implicated in the manifestations of
physiological and behavioral symptoms of anxiety and depres-
sion. Inhibition of the NK-1 receptors might cause attenuation
of the aforementioned symptoms, a hypothesis bolstered by
reports from Merck and Pfizer on their respective NK1
inhibitors.1 (2-Chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]-
* Corresponding author. E-mail: kopach_michael@lilly.com.
(1) (a) Kramer, M. S.; Cutler, N.; Feighner, J.; Shrivastava, R.; Carman,
J.; Sramek, J. J.; Reines, S. A.; Liu, G.; Snavely, D.; Wyatt-Knowles,
E.; Hale, J. J.; Mills, S. G.; MacCoss, M.; Swain, C. J.; Harrison, T.;
Hill, R. G.; Hefti, F.; Scolnick, E. M.; Cascieri, M. A.; Chicchi, G. G.;
Sadowski, S.; Williams, A. R.; Hewson, L.; Smith, D.; Carlson, E. J.;
Hargreaves, R. J.; Rupniak, N. M. J. Science. 1998, 281, 1640. (b)
Kramer, M. S.; Winokur, A.; Kelsey, J.; Preskorn, S. H.; Rothschild,
A. J.; Snavely, D.; Ghosh, K.; Ball, W. A.; Reines, S. A.; Munjack,
D.; Apter, J. T.; Cunningham, L.; Kling, M.; Bari, M.; Getson, A.;
Lee, Y. Neuropsychopharmacology 2004, 29, 385.
(2) Timpe, C.; Borghese, A.; Coffey, S. D.; Footman, P. K.; Pedersen,
S. W.; Reutzel-Edens, S. M.; Tameze, S. L.; Weber, C. WO/2005/
042515, 2005. CAN 142:469259.
10.1021/op100157q  2010 American Chemical Society
Published on Web 08/31/2010
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1229

Scheme 2. Synthesis of (2-chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]methanone, 1
achieve productive transformations, and contamination of
isolated products with TEMPO can be problematic.5 For these
reasons, there is still a need to develop and utilize similar
oxidation catalysts that operate at much lower catalyst levels.
In the course of research, we therefore developed a second-
generation process where the lithiation step was operated at -10
°C and the Anelli-Montanari oxidation was performed with 1
mol % of the highly active AZADO catalyst.
the adiabatic heat rise from the lithiation, condensation, and
quench were measured at 33, 50, and 53 °C, respectively. For
the initial 50-gal scale pilot-plant campaign, it was expected
that these exotherms could be readily controlled by the dosage
rate of process streams and managing the utilities.7 While the
heats of reaction were large compared to most conventional
processes, they were relatively low in comparison to ortho
lithation processes for similar substrates. A key requirement for
long-term commercial manufacture was to show anion stability
in the 2-5 h range which would allow for long-term operation
of this process on a 2000-gal scale.
In Situ Lithium Diisopropylamide (LDA) Preparation.
For initial evaluation of the chemistry on lab scale, a ∼0.65 M
solution of LDA (lithium diisopropylamide) in THF/hexanes
was produced by standard protocol.8 A series of ortho metalation
reactions were then run using conditions similar to what would
ultimately be employed for scaling up under cryogenic condi-
tions (Table 1). Each reaction was carried out at between -75
and -60 °C by adding a THF solution of 2-benzenesulfonylpy-
ridine, 2, to LDA.9 The results of these experiments highlight
the importance of proper LDA stoichiometry as an additional
0.05 equiv of LDA in the reaction causes formation of
considerably more bis-alcohol impurity 6 as the kinetics of the
secondary metalation become more favorable later in the
reaction pathway when alcohol 5 is the dominant species (Table
1, entry 2). Both of the scale-up reactions performed well, but
there was higher than expected levels of residual starting
Results and Discussion
Ortho metalations are well-known in the literature but can
present significant operational hazards due to the potential
formation of highly unstable benzyne intermediates.6 Conse-
quently, the thermal hazards of the target ortho lithiation reaction
were characterized by the collection of ARC data. In this case,
(3) Process mass intensity is defined as the total mass of all materials
(kg) used to produce 1 kg of active pharmaceutical ingredient (API).
The process mass intensity calculation is based on compound 1. For
leading references see: (a) Dunn, P. J.; Wells, A. S.; Williams, M. T.
Green Chemistry in the Pharmaceutical Industry, 1st ed.; Wiley-VCH:
New York, 2010. (b) Constable, D. J. C.; Curzons, A. D.; Cunningham,
V. L. Green Chem. 2002, 4, 521. (c) Constable, D. J. C.; Curzons,
A. D.; Freitas dos Santos, L. M.; Geen, G. R.; Kitteringham, J.; Smith,
P.; Hannah, R. E.; McGuire, M. A.; Webb, R. L.; Yu, M.; Hayler,
J. D.; Richardson, J. E. Green Chem. 2001, 3, 7. (d) Eissen, M.;
Hungerbuˆhler, K.; Dirks, S.; Metzger, J. Green Chem. 2003, 5, G25.
(4) (a) Backvall, J. E., Ed. Modern Oxidation Methods; Wiley-VCH:
Weinhem, 2004. (b) Zanka, A. Chem. Pharm. Bull. 2003, 51, 888.
(c) Fritz-Langhals, E. Org. Process Res. DeV. 2005, 9, 577. (d) Rossi,
F.; Corcella, F.; Caldarelli, F. S.; Heidempergher, F.; Marchionni, C.;
Auguadro, M.; Cattaneo, M.; Ceriani, L.; Visentin, G.; Ventrella, G.;
Pinciroli, V.; Ramella, G.; Candiani, I.; Bedeschi, A.; Tomasi, A.;
Kline, B. J.; Martinez, C. A.; Yazbeck, D.; Kucera, D. J. Org. Process
Res. DeV. 2008, 12, 322. (e) Hewitt, B. D. WO/1995/9516698, 1995.
CAN 123:314263.
(7) In the event of a catastrophic utilities failure the process energetics
are not sufficient to bring the reactor temperature up to the boiling
point of the solvent system (THF/hexanes) used for the lithiation
chemistry.
(8) (a) Posner, G. H.; Lentz, C. M. J. Am. Chem. Soc. 1979, 101, 934.
(b) Grieco, P. A.; Nishizawa, M.; Oguri, T.; Burke, S. D.; Marinovic,
N. J. Am. Chem. Soc. 1977, 99, 5773. (c) Enders, D.; Pieter, R.;
Renger, B.; Seebach, D. Org. Synth. 1978, 58, 113.
(9) Pre-cooling of solutions containing 2 is not recommended due to
precipitation risk.
(5) 2010 Aldrich pricing ) $6/g (25 g quantity).
(6) (a) Rawalpally, T.; Ji, Y.; Shankar, A.; Edwards, W.; Allen, J.; Jiang,
Y.; Cleary, T. P.; Pierce, M. E. Org. Process Res. DeV. 2008, 12,
1293. (b) Hickey, R. M.; Allwein, S. P.; Nelson, T. D.; Kress, M. H.;
Sudah, O. S.; Moment, A. J.; Rodgers, S. D.; Kaba, M.; Fernandez,
P. Org. Process Res. DeV. 2005, 9, 764. (c) Bridges, A. J.; Lee, A.;
Maduakor, E. C.; Schwartz, C. E. Tetrahedron Lett. 1992, 33 (49),
7495.
1230
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Table 1. Ortho metalation reactions with 1.05 equiv in situ
Table 2. Evaluation of commercial vs in situ LDA
prepared LDAb
LDA
entry (equiv)
LDA
Source
conversion remaining conversion
entry
scale (g)
5 (%)
2 (%)
6 (%)
to 5 (%)
2 (%)
to 6 (%)
1
2
3
4
5
6
5
5a
93.8
83.3
88.3
88.4
92.4
91.2
2.1
5.7
7.0
8.0
3.2
3.7
2.1
6.3
1.3
1.4
1.3
2.5
1
1.05 in situ
85.8
80.8
69.9
66.2
82.4
82.7
76.2
48.8
11.7
87.7
7.1
6.5
12.1
10.8
5.5
6.2
6.5
15.8
25.4
3.6
0.69
3.4
2
1.10 in situ
50
100
5
3
1.30 in situ
7.9
4
1.00 commercial
1.10 commercial
1.15 commercial
1.20 commercial
1.2
5
1.7
5
6
1.6
a 1.10 equiv of LDA. b All HPLC data reported in tables are uncorrected area
% at 250 nm.
7
2.9
8
2.0
3.0
commercial
commercial
8.2
9
12.7
2.4
10a
1.05 commercial
material 2 and low levels of the bis-product 6; it was suspected
that this was due to an undercharge of LDA.10
a 500 g scale run.
As we prepared to adapt this chemistry to 50-gal scale, the
lithiation/condensation operation temperature was recentered at
-50 °C due to commercial manufacturing constraints. Process
modeling data showed that by assuming a reactor temperature
of -50 °C, a jacket temperature of -60 °C, and an estimated
heat of reaction of 100 kJ/mol for lithiation of 2, the minimum
addition time on a 2000-gal scale would be about 3 h.
Consequently, one of the goals of the pilot-plant campaign was
to evaluate longer addition times to see if there were any adverse
effects to the process. Unfortunately, precooling the THF
solution containing 2 was not an option because the solid would
precipitate.
Evaluation of Commercial LDA. The use of n-butyllithium
solutions on large scales is replete with hazards and requires
specialized handling procedures.11 In fact, all formulations of
n-butyllithium have been classified as pyrophoric by the DOT.
A further hazard associated with direct use of n-butyllithium is
liberation of butane during processing, which requires appropri-
ate venting controls. A potentially attractive option is the use
of commercial LDA, which is significantly less hazardous,12
and is available on large scale from FMC Lithium as a 2.0 M
solution in a mixture of THF/heptanes/ethylbenzene. However,
there are some known liabilities of commercial LDA as these
solutions undergo decomposition to LiH and the corresponding
imine, the rate of which is primarily impacted by storage
temperature. Thus, it is not surprising that there are several
reports of inconsistent performance with commercial LDA,
especially when an aged source is used.13 In comparison,
n-butyllithium solutions are stable at ambient temperature, but
are much more sensitive to moisture-based degradation. How-
ever, in our case the overall improved safety profile with
commercial LDA was a significant driver to thoroughly evaluate
that option. A fresh supply of commercial LDA was evaluated
head to head with in situ prepared LDA (Table 2). The
commercial LDA was diluted with THF to produce a 0.65 M
solution comparable in concentration to the in situ produced
LDA.
The results achieved using commercial LDA were similar,
although not identical, to using in situ prepared LDA and
demonstrated a close correlation between LDA stoichiometry
and the amount of bis-addition product 6 (Table 2). With these
results in hand, commercial LDA was evaluated on a 500 g
scale, and the results showed excellent conversion albeit with
elevated levels of bis-byproduct 6 compared to when in situ
LDA was employed (Table 2, entry 10 vs entry 1). To overcome
solubility issues we also evaluated an inverse addition protocol
at double the normal dilution which gave very similar results
to the standard addition mode. Anion stability was tested at
-50 °C, by aging the mixture for 5 h prior to quenching with
aldehyde 4, which produced one of the best results to date. The
stoichiometry of aldehyde 4 was shown to be not as crucial as
the LDA charge since excess aldehyde was oxidized to
2-chlorobenzoic acid which was removed in the aqueous
workup. For standard conditions, a slight excess of aldehyde
(1.05 equiv) was employed because if there were elevated levels
of bis-anion present, this would also result in unreacted starting
material. With the data showing that fresh commercial LDA
performs as well as in situ generated LDA when scaled, the
(10) In this case, the scale-up runs were performed with n-butyllithium
which had been used for several prior experiments.
(11) (a) Dezmelyk, E. W.; Reed, R. S. Ind. Eng. Chem. 1961, 53 (6), 68A.
(b) Schwindeman, J. A.; Woltermann, C. J.; Letchford, R. J. Chem.
Health Saf. 2002, (May/June), 6.
(13) (a) Smith, I. H.; Alorati, A.; Frampton, G.; Jones, S.; O’Rourke, J.;
Woods, M. W. Org. Process Res. DeV. 2003, 7, 1034. (b) Collum,
D. B.; Singh, J. K.; Hoepkar, P. C.; Gupta, L. J. Org. Chem. 2009, 74
(5), 2231. (c) Morrison, R. C.; Hall, R. W.; Rathman. T. L. U.S. Pat.
4,595,779, 1986.
(12) Commercial LDA is nonpyrophoric, as determined by the Department
of Transportation (DOT). The official test for pyrophoricity is detailed
in the DOT regulations. Code of Federal Regulations, 49 CFR 173,
Appendix E.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1231

Table 3. Pilot-plant ortho lithiation/condensation reaction
Scheme 3. Standard condition for Anelli-Montanari
oxidation
was observed with only a mixture of starting materials (2 and
4) observed. Thus, it was evident that the process could be
carried out at a much higher temperature, thereby eliminating
the need for extreme cryogenic conditions. The next step was
to determine whether the anion of 3 could be generated at
warmer temperatures. Thus, generation of the LDA was carried
out at -10 °C, followed by addition of 2 and quenching with
2-chlorobenzaldehyde. The crude product was formed with 87%
conversion and barely detectable levels of bis-addition product
6. From this data it was surmised that the bis-anion was
significantly less stable at warmer temperature. While the
quantity of remaining starting material was higher than desired,
there was not a considerable difference between this temperature
and some of the lower-temperature reactions (cf. -50 °C, Table
4, entry 2).
addition addition
time of time of LDA conversion remaining conversion
batch 2 (min) 3 (min) (equiv) to 5 (%)
2 (%)
to 6 (%)
1
2
3
4
5
27
21
58
40
39
60
1.05
1.15
1.12
1.10
1.12
78
75
78
77
87
9.9
7.6
4.7
7.8
4.5
2.6
4.6
5.0
6.8
4.2
117
17
120
27
Table 4. Anion stability study
Oxidation Reaction Development. With improved condi-
tions developed for the lithiation/condensation step, our attention
was directed toward the oxidation step. The crude alcohol/THF
solution was initially solvent-exchanged to dichloromethane,
which is the most prevalent solvent for the Anelli-Montanari
oxidation.14 Initial experiments used 0.07 equiv of TEMPO and
3.5 equiv of 12 wt % aqueous sodium hypochlorite in a solution
of dichloromethane and crude alcohol 1 (Scheme 3). These
conditions gave complete conversions without formation of
impurities other than those indigenous to the ortho metalation
reaction such as ketone 7. Sodium bicarbonate was used to
lower the pH of the bleach to 8.5-9.5 and to buffer any acid
formed in the reaction, because sodium hypochlorite is unstable
at a lower pH. Additional water generally does not affect the
reaction rate of the biphasic system, and oxidations are usually
complete within 1 h. We also examined two literature alterna-
tives to the current process. The first method, employing
trichloroisocyanuric acid, potassium bromide, and TEMPO in
acetone and water delivered only a 2:1 ratio of starting material
to product after 21 h, even after heating to 50 °C for the last
2 h. The second method, using sodium hypochlorite and acetic
acid, showed only 1.7% conversion to product after 19 h; thus,
both of these approaches were abandoned.
reaction
entry temperature (°C)
conversion remaining conversion
to 5 (%)
2 (%)
to 6 (%)
1
2
3
4
5
6
-70
-50
-40
-20
-10
0
89
5.7
11
6.8
8.0
6.8
10.3
1.0
1.0
0.1
0.8
0.6
0.7
87
93
90
92.1
88.0
decision was made to move ahead with commercial LDA for
the pilot plant campaign. A total of five lots were run in the
pilot plant with each using a 3.5 kg charge basis of 2-benz-
enesulfonyl pyridine 2, the crude product solutions were
quenched with 3 N HCl, and then carried directly into the
oxidation and crystallization steps for production of 1 (Table
3). As expected, the process streams with the faster feed rates
resulted in improved conversion and purity, but notably all feed
rates examined produced satisfactory results.
Second-Generation Lithiation/Condensation Studies.
While the results from the pilot-plant scale lithiation/condensa-
tion reaction were acceptable, an attractive green chemistry
improvement would be to operate this chemistry at warmer
temperatures. With this objective in mind, a significant goal of
this new study was to ascertain the stability of the anion 3 at
milder temperatures. Thus, the initial study on 2-g scale involved
preparation of the LDA at -70 °C and addition of 2, followed
by adjustment to the desired temperature (Table 4). The anion
solution was then held for 2 h at the selected temperature,
followed by addition of aldehyde 4.
Since 1 is freely soluble in dichloromethane, a greener
solvent was desired for crystallization of 1. The result of
solubility studies revealed that 1 was sparingly soluble in toluene
(14 mg/mL) which made it a potentially attractive crystallization
solvent. Ultimately, the goal was to replace dichloromethane
entirely in both the reaction and isolation. For the oxidation
reaction we found that toluene performed equivalently to
dichloromethane, and as the reaction progressed in toluene,
compound 1 precipitated from the reaction and could be isolated
The results of the anion study revealed considerable stability,
even up to 0 °C. However, when the anion temperature was
raised further to room temperature, degradation of the anion
(14) (a) Anelli, P. L.; Biffi, C.; Montanari, F.; Quici, S. J. Org. Chem.
1987, 52, 2559. (b) Anelli, P. L.; Biffi, C.; Montanari, F.; Quici, S. J.
Org. Chem. 1989, 54, 2970.
1232
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Table 5. Reaction solvent and crystallization screen for Anelli-Montanari oxidation
entry rxn solvent
xtal solvent
crude 1 (%) crude 7 (%) yield 1 (%) isolated purity 1 (%) isolated 7 (%) in product
a1
a2
a3
4
CH2Cl2
toluene
EtOAc
CH2Cl2
EtOAc
EtOAc
EtOAc/heptane
toluene
EtOAc/heptane
EtOAc/heptane
EtOAc/heptane
EtOAc/heptane
98
98
99
91
78
90
0.3
n.d.
n.d
2.0
1.5
2.4
90
86
90
85
84
90
99.9
99.8
n.d.
n.d.
n.d.
0.43
0.39
1.2
100
99.1
99.0
97.2
5
b6
a Purified 5 used. b 500-g scale-up.
Table 6. Anelli-Montanari oxidation temperature evaluation
entry
start temp (°C)
end temp (°C)
addition time (min)
remaining 5 (%)
reaction conversion to 1 (%)
1
2
3
4
5
6
2
7
10
15
22
23
16
17
21
27
36
38
25
20
20
15
30
30
0.3
n.d.
0.9
0.8
2.1
2.3
79
89
86
78
79
83
without additional workup. However, the reaction mixture
became very thick as the product precipitated, and we were
concerned that poor mass transfer could negatively affect
completion as the process was scaled. Additionally, filtration
of the biphasic mixture to isolate 1 was found to be very slow
on lab scale which would not translate well to plant scale.
Additional solvent screens were run for the reaction and
crystallization (Table 5).
The process screen revealed that ethyl acetate was an
effective solvent for both reaction and crystallization; with
purified alcohol 5 as a substrate a 99% conversion was observed
within an hour, and an isolated yield of 90% was achieved. As
the reaction progressed in ethyl acetate, the product precipitated.
However, the isolated yield was about 10% lower due to loss
of product to filtrate. Heptane was found to be an effective
antisolvent, and after the reaction was complete in EtOAc, the
mixture was warmed to ambient temperature and 8 volumes of
heptanes were added. Under these conditions it was possible
to telescope crude alcohol 5 directly to 1 in 84-90% yield with
97-99% chemical purity (Table 5, entries 5, 6). In contrast to
the toluene crystallization, the biphasic filtration in the ethyl
acetate/heptane/water system was exceptionally rapid, occurring
within a few minutes. With respectable small-scale results in
hand, we tested the telescoped process on a 500-g scale of
2-benzenesulfonylpyridine 2. The ortho metalation proceeded
as expected, with 3.6% residual starting material and 2.4% bis-
impurity 6 (Table 2, entry 10). The oxidation reaction also ran
as expected with less than 1% alcohol 5 remaining after one
hour and the isolated yield (90%) was the best to date; however,
the purity of the isolated solid was lower than expected with
2.8% total impurities including 1.2% bis-adduct 7 (Table 5,
entry 6). We were able to reslurry the product in 1:1 ethyl
acetate/heptane and recovered 94% of 1 with 99.4% purity. This
indicated that the initial filter cake wash might not have been
very effective, possibly due to channeling.
During the oxidation reaction we observed that the addition
of bleach was exothermic; therefore, the reaction contents are
usually cooled before the addition. Throughout most of the
development phase, we cooled the alcohol 5 solution to 0-5
°C which allowed rapid bleach addition while maintaining a
pot temperature of <15 °C on small scale. To potentially save
on heating and cooling time on production scale, we evaluated
warmer temperatures for the oxidation (Table 6). The amount
of remaining 5 could be driven below 1% except when the
oxidation process was started at room temperature and allowed
to exotherm to 35-40 °C (Table 6, entries 5, 6). Most likely,
the decreased conversion is due to NaOCl degradation which
is common at higher temperatures.
Both of the reactions that commenced between 10-15 °C
showed a steady exotherm from the outset of the bleach feed,
with maximum end temperatures of 21-27 °C (Table 6, entries
3, 4). With the success of these experiments, we decided that
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1233

Table 7. Telescoped pilot-plant synthesis of 1
yield (kg) yield (%)
purity 1
2 (%)
7 (%)
entry conversion to 1 (%) remaining 2 (%) conversion to 6 (%)
isolated
isolated
(%) isolated in product in product
1
72
72
76
75
82
75
10.3
8.7
5.0
6.4
3.9
6.9
2.8
7.1
5.2
5.1
4.2
4.9
4.4
76
84
88
87
92
85
98.2
94.1
96.8
93.7
96.4
95.8
3.4
2.8
1.4
1.1
0.6
1.9
1.2
6.1
3.1
5.0
3.3
3.7
2
4.8
3
5.1
4
5.0
5
5.2
ave
24.5 (net)
Table 8. Streamlined process and direct isolation of 1 from EtOAc
entry scale 2 (g) crude 1 (%) crude 2 (%) crude 7 (%) yield (%) isolated purity 1 (%) isolated 2 (%) in product 7 (%) in product
1
2
3
4
5
5
50
100
400
400
83
82
83
92
90
4.6
5.1
6.2
1.9
2.2
5.8
1.5
1.1
2.3
2.7
75
82
81
87
85
98.2
99.0
99.2
97.8
97.7
0.50
0.50
0.57
0.67
0.91
0.91
0.21
0.24
0.85
0.94
the oxidation step in the pilot plant would start at ∼15 °C and
the exotherm would be limited to not more than 20 °C by jacket
temperature control. Prior to the start of the pilot-plant campaign
we also evaluated a reduced TEMPO catalyst load from the
standard level of 7 mol %. Unfortunately, reduction of the
catalyst level to 3 mol % resulted in only a ∼50% conversion,
and addition of kicker charges of TEMPO did not increase the
conversion, so we decided to stay with the standard catalyst
load of 7 mol %.
Pilot-Plant Synthesis of 1. The telescoped process for the
500-g scale synthesis of 1 was run in the pilot plant on a 50-
gal scale. A total of 25 kg of 1 were produced in five lots with
average 95.8% purity (Table 7). The initial run resulted in a
high level of residual starting material 2 (10.3%) that gave
correspondingly higher amounts in the isolated product. The
LDA stoichiometry was increased in the next lot and the bis-
byproduct 6 was significantly elevated (Table 7, entry 2). The
Anelli-Montanari oxidation performed as expected, efficiently
converting crude alcohol 5 and bis-alcohol impurity 6 to the
desired product 1 and bis-diketone 7 within one hour. The main
difference between the pilot-plant runs and supporting-labora-
tory runs was that the oxidation step was run at a warmer
starting temperature of 14 °C, and an exotherm to 20 °C was
observed during the bleach addition. The main issue encountered
was variable impurity levels on pilot scale, with respect to the
bis-adduct 7. Fortunately, while the chemical purity of 1 was
low for an API starting material, a rework was not necessary.
Forward-processing of 1 to the penultimate intermediate
involves a SNAr reaction that converts the bis-ketone 7 to a
byproduct that is completely removed during the workup. Thus,
all lots of 1 produced in this campaign were converted to the
penultimate intermediate, none of which contained any impuri-
ties related to the purity profile of 1.
Second-Generation Studies. Although the yield goals were
achieved for the pilot-plant campaign, the higher than expected
impurity levels and process variability needed to be addressed
before outsourcing the process to contract manufacturing
organizations. After the pilot-plant campaign was complete, a
survey of methods to purify 1 was performed. We found that
bis-impurity 7 could be reduced by a factor of 5 with a 92%
recovery with an ethyl acetate reslurry at 0-5 °C. Numerous
other solvent systems were studied, but ethyl acetate was by
far the most effective. With this data in hand, the next logical
step was to evaluate heptane removal from the precipitation/
isolation step in the standard process. For these studies we used
1.05 equiv of in situ generated LDA, and series of experiments
were run while progressively increasing the scale (Table 8).
The final two experiments were performed on a 400-g scale of
2 in 22 L equipment. In both cases, after the reaction was
complete the mixture was cooled to 0-5 °C and the entire
contents of the biphasic mixture were filtered to isolate 1 without
the addition of heptanes. Under these conditions, the two 400-g
scale-up runs averaged 86% yield of isolated product with
97.8% chemical purity, a 2-3% improvement in purity from
that in the pilot-plant campaign. Analysis of the filtrate revealed
a 6-7% soluble product loss which was about 4-5% worse
than filtrate loss observed in the pilot plant. The overall mass
yield of 1 for the streamlined process was the same as for the
1234
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

the preparation of 1 from 2. Experiments were conducted using
∼2 g of 2 in which either 0.07 equiv of TEMPO, 4-MeO-
TEMPO, or 4-acetamido-TEMPO or 0.01 equiv of AZADO
were employed (Table 10). As expected, similar yields of 1
were achieved on employing a 0.01 equiv loading of AZADO
in comparison to a 0.07 equiv loading of TEMPO.
Figure 1. TEMPO alternatives.16
Table 9. Catalyst study in the oxidation of 5 to 1
Borax (sodium tetraborate, Na2B4O7) has been described as
an environmentally friendly alternative to KBr by Augustine.17
Two experiments were conducted using 0.6 equiv of borax as
a 2% aqueous solution. With 7 mol % TEMPO this afforded 1
in 74% yield, while the use of 1 mol % of AZADO and borax
gave 1 in 86% yield. However, the isolated ketone 1 contained
borax as an impurity which coprecipitated from the reaction
medium, making this a less attractive option. The final refine-
ment we evaluated was performing the fully telescoped process
with 1 mol % AZADO where the lithiation step was operated
at -10 °C. In this case the process was scaled up to 5 L scale
(using 200 g of 2) with an overall yield of 71% being achieved
(Scheme 4). With further optimization it is anticipated that the
yield of the process could match the TEMPO-based process.18
The AZADO catalyst is presently available in small quantities
from Aldrich (250 mg @ $187.50). While bulk quantities for
commercial scale do not appear available without custom
manufacture, it is hoped that the successes reported here and
elsewhere will produce a commercial market for the highly
active AZADO catalyst.19
Process Impurities. Both 2-benzenesulfonyl pyridine 2 and
alcohol 5 can be found in varying amounts in isolated 1 due to
incomplete reactions. The synthesis of 2 for all studies was
prepared in high yield and purity by the Trankopach reaction.20
The principal process impurity observed was bis-ketone 7 which
was formed by the reaction of monolithiated 2 with excess LDA
followed by the subsequent reactions. The dilithiated species
where the second lithiation occurs on the other benzene ring
then reacts with 2-chlorobenzaldehyde to produce diol 6 that
undergoes oxidation to the diketone 7. The bis-adduct was found
to be significantly more insoluble in water than 1 and of interest;
secondary reactions with the parent pyridine ring were not
observed. Compounds 8 and 9 are positional isomers of 1 and
can be traced back to contamination of 2-chlorobenzaldehyde
with the 3- and 4-chloro isomers (Figure 2). We prepared the
3- and 4-chloro isomers of 1 without any significant issues using
the telescoped TEMPO process, and both analogues behaved
similarly to the 2-chloro isomer. These standards were used to
show that production batches of 1 typically contained non-
detectable levels of isomers 8 and 9.
yield of 1 (%)a
loading
amount
(equiv) TEMPO 4-MeO-TEMPO 4-AA-TEMPO AZADO
0.07
0.01
0.005
93
53
98
62
54
90
47
46
94
99
97
25b
a Combined yield of product, which precipitated, plus extraction of organic
layer. b No precipitate.
pilot-plant campaign. However, the potency-corrected yield for
the second-generation approach was ∼3% higher due to
improved purity of the product. The overall improvements in
yield and quality appear to be mainly due to higher-purity ortho
lithiation/condensation reactions. Most notably, bis-impurity 7
in all experiments was reduced to <1% by the crystallization.
The streamlined process can not only accommodate a wider
variability in the ortho lithiation reaction but also significantly
reduces the process mass intensity by 25%. Ultimately, these
conditions will be recommended to contract manufacturing
organizations targeting the production of 1.
Alternative Oxidant Screens. Despite the successes of the
TEMPO system, an opportunity existed to investigate alterna-
tives with improved catalytic activity (Figure 1). TEMPO,
4-acetamido-TEMPO, 4-methoxy-TEMPO, and AZADO15 were
evaluated in the sodium hypochlorite oxidation of 5 to 1 (Table
9).
At loadings of 0.07 equiv, the four organocatalysts exhibited
similar catalytic activity. On decreasing the loading to 0.01
equiv, the efficiency of the TEMPO-based catalyst was reduced,
while AZADO maintained its high activity. When the catalyst
loading was decreased to 0.005 equiv, the efficiency of the
TEMPO-derived catalysts were further reduced (with TEMPO
proving to be the least efficient), while a 97% conversion was
still achieved with AZADO. However, at this lower loading of
AZADO less product precipitated out of the reaction solution
(76% vs 94-99%) than with the corresponding experiments
which were conducted using the higher catalyst loadings, and
on extracting the mother liquor a further 21% of 1 was isolated.
In this case the crystallization mixture was ∼20% more dilute
which likely contributed to the higher loss of yield of product
to filtrate. While the experiment was not repeated with the
typical substrate concentration, we decided to continue further
experimentation with 0.01 equiv AZADO. Enhanced reactivity
of AZADO has been postulated to be due to reduced steric
hindrance near the nitroxyl group (Figure 1).13 These same four
oxidation catalysts were then used in the two-step process for
(15) Shibuya, M.; Tomizawa, I.; Suzuki, I.; Iwabuchi, Y. J. Am. Chem.
Soc. 2006, 126, 8412.
(16) Criminna, R.; Pagliaro, M. Org. Process Res. DeV. 2010, 14, 245.
(17) (a) Prakash, I.; Tanielyan, S. K.; Augustine, R. L.; Furlong, K. E.;
Scherm, R. C.; Jackson, H. E. U.S. Pat. 6,825,384, 2004. CAN 142:
6014. (b) Tanielyan, S. K.; Augustine, R. L.; Prakash, I.; Furlong,
K. E.; Scherm, R. C.; Jackson, H. E. WO/2004/067484, 2004. CAN
141: 190314. (c) Tanielyan, S.; Augustine, R.; Meyer, O.; Korell, M.
U.S. Pat. 7,030,279, 2006. CAN 144: 393182.
(18) Preliminary experiments indicate that 2-methyl THF may be a viable
alternative to THF for the lithiation chemistry.
(19) When running the oxidation step in the clinical pilot plant in 2005,
AZADO was not available.
(20) Kopach, M. E.; Trankle, W. G. Org. Process Res. DeV. 2007, 11,
913.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1235

Table 10. Variation of catalyst in the synthesis of 1 from 2
catalyst
loading (equiv)
yield (%)
TEMPO
0.07
0.07
0.07
0.01
74-80
63
4-methoxy-TEMPO
4-acetamido-TEMPO
AZADO
69
78
Scheme 4. Preparation of the ketone 1 using AZADO methodology
Conclusions
flow rate of 1.0 mL/min, a wavelength of 250 nm, at a
temperature of 23 °C. Mobile phases (A and B) were 0.1%
H3PO4 in H2O and acetonitrile, respectively. HPLC retention
times were as follows: compound 1 ) 11.2 min; compound 2
) 9.3 min; compound 4 ) 10.5 min; compound 5 ) 10.7 min;
compound 6 ) 11.8 min; compound 7 ) 12.2 min; compound
8 ) 11.6 min; compound 9 ) 11.6 min.
A highly selective ortho lithiation/condensation/oxidation
telescoped process to produce 25 kg of 1 was demonstrated in
our pilot plant. Commercial LDA was used for the lithiation
step which operated at -50 °C. The efficient Anelli-Montanari
oxidation using catalytic TEMPO (7 mol %) was used to oxidize
the intermediate ketone to 1. A second-generation process was
also developed which delivered the target compound in
improved purity using ethyl acetate as both oxidation and
crystallization solvent. Significantly, this improved process had
a 25% reduced process mass intensity relative to that of the
piloted route. In addition, we developed an alternative process
option where the lithiation chemistry can be performed at -10
°C and the Anelli-Montanari oxidation can be run with the
highly efficient AZADO catalyst.
(2-Chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]metha-
none, 1. Preparation A: Pilot-Plant Synthesis with Commercial
LDA. Tetrahydrofuran (17 L) was charged to a 30-gal glass-
lined cryo-rated vessel and cooled to -55 °C. Commercial LDA
(2.1 M, 6.8 kg, 8.4 L) was charged and the vessel temperature
readjusted to -55 °C. A solution of compound 1 (3.5 kg) in
THF (11 L) was added over 17 min at <-50 °C. The mixture
was stirred for 2.5 h at -55 °C, and then a solution of aldehyde
4 (2.39 kg) in THF (4 L) was added over 27 min at <-50 °C.
The solution was stirred at -50 °C for 2 h at which time HPLC
analysis showed 4.5% of compound 2 remaining. Aqueous 3
N HCl (29.4 kg) was added to the reaction over 13 min while
allowing the contents to warm to 10 °C. The contents were
warmed to 20 °C, and the layers were allowed to separate. The
lower aqueous layer was back extracted with ethyl acetate (21
L), the combined organic layers were washed with 8.5 wt %
NaHCO3 (25 kg), and the resulting organic solution was vacuum
distilled with a maximum jacket temperature of 55 °C to a
volume of 15-20 L. Ethyl acetate (35 L) was added, and the
solution was reconcentrated to a volume of 15-20 L. Ethyl
Experimental Section
General. For non-pilot-plant and kilo-lab operations solvents
were distilled prior to use as follows. Tetrahydrofuran was
freshly distilled over sodium benzophenone ketal; ethanol was
distilled from magnesium ethoxide and stored over activated 3
Å molecular sieves; ethyl acetate was distilled over potassium
carbonate; diisopropylamine was freshly distilled over calcium
hydride. High performance liquid chromatography (HPLC)
analysis was performed on a Waters Alliance 2690 separations
module with a Waters 486 tunable absorbance detector using a
Zorbax RX-C8 (25 mm × 4.6 mm × 5 mm) column using a
Figure 2. Impurities identified in the preparation of 1.
1236
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

acetate (15 L), aqueous 10% potassium bromide solution (11.2
kg), and 2,2,6,6-tetramethyl-1-piperidine-N-oxide (TEMPO)
(180 g) were added to the 50-gal vessel, and the mixture was
cooled to 15 °C. A solution of 12% aqueous sodium hypochlo-
rite (31 L), sodium bicarbonate (1.7 kg), and water (18 kg) was
added to the biphasic mixture over 45 min, while allowing the
reaction temperature to heat to 20 °C. After the reaction mixture
stirred for 1 h at 20-25 °C, HPLC analysis of an aliquot showed
<1% of compound 5 remaining. Heptane (28 L) was added to
the product mixture over 30 min. The mixture was then aged
for 3 h at 20 °C, the slurry was filtered, and the cake was washed
with heptanes (10 L), followed by water (10 L) and then
heptanes (10 L). The product 1 was dried under vacuum at 55
°C for 44 h to produce 5.2 kg (92% yield) of 1 with a potency
of 96.4%. νmax/cm-1 (KBr) 3445, 1682, 1298, 1159; δH (400
MHz, CDCl3) 7.39 (1H, ddd, J 7.8, 7.2, 1.6), 7.45-7.56 (5H,
m), 7.59-7.65 (1H, m), 7.74 (1H, dd, J 7.6, 1.6), 7.81 (1H,
dd, J 8.0, 1.6), 7.98-8.04 (2H, m), 8.78 (1H, dd, J 4.4, 1.6);
δC (75.5 MHz, CDCl3) 126.4, 126.8, 128.9, 129.2, 131.5, 133.1,
133.8, 133.9 (8 × CH), 134.2, 134.8 (2 × C), 136.7 (CH),
137.1, 138.6 (2 × C), 150.5 (CH), 155.1, 191.5 (2 × C). HRMS
(ES+) Exact mass calculated for C18H1335ClNO3S [(M + H)+],
358.0305. Found 358.0298, m/z (ES+) 359.9 {[(C18H1337Cl-
NO3S) + H+], 40%}, 358.0 {[(C18H1335ClNO3S) + H+],
100%}. Process Mass intensity ) 53.4 kg/kg 1.
Preparation B: Second-Generation Kilo-Lab Process with
EtOAc Crystallization. Tetrahydrofuran (2.0 L) and diisopro-
pylamine (335 mL) at -75 °C were treated with 1.6 M
n-butyllithium in hexanes (1.28 L) over 30 min at <-60 °C.
The resulting light-yellow solution was stirred at -60 to -70
°C for an additional 30 min; then a solution of compound 1
(400.5 g) in THF (1.2 L) was added over 20 min at <-60 °C.
A greenish-yellow suspension formed during the addition, and
the mixture was stirred for 2.5 h at -60 °C. A solution of
compound 4 (272.3 g) in THF (400 mL) was added over 30
min at <-60 °C. The solution was stirred at -60 °C for 1 h at
which time HPLC analysis showed 3.6% of compound 1
remaining. Aqueous 3 N HCl (2.8 L) was added over 30 min
while allowing the contents to warm to 2 °C (the cooling bath
was removed about halfway through the addition). The contents
were stirred until the temperature warmed to 8 °C over 30 min.
Stirring was stopped, the layers were separated, and the lower
aqueous layer was extracted with ethyl acetate (4 L). The
combined organic layers were stirred with saturated sodium
bicarbonate solution (1 L) for 15 min, and the layers were
allowed to separate. The organic solution was vacuum distilled
(bath temperature was 48 °C) to a volume of ∼1 L. Ethyl acetate
(4 L) was added, and the solution was reconcentrated to a
volume of 2 L. Ethyl acetate (2 L), aqueous 10% potassium
bromide solution (1.3 L), and then 2,2,6,6-tetramethyl-1-
piperidine-N-oxide (TEMPO) (20.8 g) were added, and the
mixture was cooled to 3 °C. A mixture of 12% aqueous sodium
hypochlorite (3.6 L), sodium bicarbonate (192.2 g), and water
(2.2 L) was added to the biphasic solution over 1 h while
allowing the reaction temperature to rise to 22 °C. After stirring
the reaction mixture for 1 h at 20-25 °C, HPLC analysis of an
aliquot showed 0.73% of compound 5 remaining. The mixture
was cooled to 0 °C and stirred for 3 h. The slurry was filtered
through polypropylene cloth on a 24 cm stainless steel single-
plate filter (the entire filtration took 50 s), and the filter cake
was washed with cold ethyl acetate (2 L) followed by water (2
L). The product, 1 was dried under vacuum at 45 °C overnight
to constant weight. The quantity of compound 1 produced was
571.8 g (87% yield) with a potency of 97.8%.
Preparation C: In Situ Prepared LDA at -10 °C with
AZADO Oxidation. Tetrahydrofuran (1.0 L) and diisopropyl-
amine (169 mL, 1197.8 mmol, 1.3 equiv) were cooled to -10
°C, and 1.75 M n-butyllithium (553 mL, 967.5 mmol, 1.05
equiv) was added over 30 min at -10 to 0 °C. The reaction
mixture was stirred for 0.5 h at -10 °C. A solution of 2 (202.0
g, 921.4 mmol, 1.0 equiv) in THF (600 mL) was added over
20 min, at -10 and 0 °C. The mixture was then stirred at -10
°C for 2 h. 2-Chlorobenzaldehyde (110 mL, 977 mmol, 1.15
equiv) was added over 15 min. The mixture was stirred for 1 h
at -10 °C. Aqueous 3 N HCl (1.4 L) was added over 30 min
as the reaction mixture was allowed to warm to ∼10 °C. The
layers were separated. The aqueous layer was extracted with
ethyl acetate (2.0 L), and the combined organic layers were
stirred with saturated aqueous sodium bicarbonate (2.0 L) for
15 min. The layers were separated, and the organic layer was
washed with brine (2.0 L), dried with MgSO4, and concentrated
under reduced pressure to a volume of 1.0 L. Ethyl acetate (1.0
L) was added, and aqueous KBr (10% w/w) solution (660 mL)
was added, followed by AZADO (1.4025 g, 9.2 mmol), and
the reaction mixture was cooled to 0 °C. A mixture of 9.3 wt
% aqueous sodium hypochlorite (2.58 kg), sodium bicarbonate
(123.85 g, 1474.2 mmol), and water (750 mL) was added over
1 h while allowing the reaction mixture to warm to room
temperature. Following stirring at room temperature for 3 h, a
thick precipitate was observed. The mixture was cooled to -5
°C and maintained at this temperature for 2 h. The resultant
slurry was filtered through a sintered glass funnel, and the cake
was rinsed with cold ethyl acetate (500 mL), water (500 mL),
and ethyl acetate (500 mL) to yield the ketone 1 (233.66 g,
71%) as a white solid with a purity of 98.2%.
(2-Chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]metha-
nol, 5. A solution of sodium borohydride (0.91 g, 24.1 mmol)
in ethanol (50 mL) was added over 15 min to a solution of the
ketone 1 (8.48 g, 23.7 mmol) and ethanol (250 mL) at 0 °C.
The reaction mixture was allowed to warm to room temperature
and stirred overnight. Water (100 mL) was added with stirring
to quench the reaction, and the crude reaction mixture was
concentrated under reduced pressure to yield a white aqueous
residue. Ether (100 mL) was added, and the layers were
separated. The aqueous layer was washed with ether (2 × 100
mL), and the combined organic layers were washed with brine
(100 mL), dried, filtered, and concentrated under reduced
pressure to give the alcohol 5 (5.64 g) as a white solid.
Purification (of 2 g) by recrystallisation at 65 °C from a mixture
of ethyl acetate and tert-butyl methyl ether (added at 40 °C)
gave 5 (1.76 g) as a white solid, mp 135-138 °C; (Found: C,
59.74; H, 3.87; N, 3.67; S, 9.09; Cl, 10.03. C18H14ClNO3S
requires C, 60.08; H, 3.92; N, 3.89; S, 8.91; Cl, 9.85%); νmax
/
cm-1 (KBr) 3547, 3050, 1442, 1293, 1153; δH (300 MHz) 4.27
(1H, br s, OH, exchanges with D2O), 6.87 (1H, s, CHOH),
7.25-7.35 (4H, m), 7.41 (1H, ddd, J 7.8, 6.2, 2.6), 7.57 (2H,
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1237

2 × overlapping dd J 7.9, 7.4), 7.67 (1H, ddd, J 7.4, 4.4, 1.0),
7.84 (1H, d, J 8.0), 8.05-8.11 (2H, 2 × overlapping d, J 8.6),
8.53 (1H, dd, J 4.3, 1.8); δC (75.5 MHz) 66.1 (CH, CHOH),
126.0, 126.3, 127.5, 128.00, 128.01, 128.2, 128.5 (7 × CH),
131.0 (C), 132.9, 136.7 (2 × CH), 137.6, 137.7 (2 × C), 147.2
(CH), 155.4 (C); HRMS (ES+) Exact mass calculated for
C18H1535ClNO3S [(M + H)+], 360.0462. Found 360.0465, m/z
(ES+) 362.0 {[(C18H1437ClNO3S) + H+], 33%}, 360.0
{[(C18H1435ClNO3S) + H+], 100%}.
(3-Chlorophenyl)[2-(phenylsulfonyl)pyridin-3-yl]metha-
none, 8. The general procedure described for the preparation
of 1 was used, except 3-chlorobenzaldehyde was used in place
of 2-chlorobenzaldehyde for the condensation reaction. These
conditions produced compound 8 as a tan solid in 72% yield
with 94% purity. Purification of compound 8 by flash chroma-
tography on silica gel using 20% cyclohexane in tert-butyl
methyl ether produced a white solid in 97% purity, mp 147-150
°C; νmax/cm-1 (KBr) 3081, 1677, 1573, 1430, 1322; δH (300
MHz, CDCl3) 7.45 (1H, t, J 7.9), 7.50-7.71 (6H, m), 7.74 (1H,
dd, J 7.8, 1.6), 7.80 (1H, dd, appears as overlapping t, J 1.9,
1.8), 7.97-8.06 (2H, m), 8.83 (1H, J 4.6, 1.6); δC (75.5 MHz,
CDCl3) 126.5, 128.2, 129.2, 129.3, 129.6, 130.2, 134.0, 134.1
(8 × CH), 135.2, 135.3 (2 × C), 136.8 (CH), 138.1, 138.5 (2
× C), 151.0 (CH), 155.6 (C), 196.9 (C); HRMS (ES+): Exact
mass calculated for C18H1335ClNO3S [(M + H)+], 358.0305.
Found 358.0300, m/z (ES+) 359.9 {[(C18H1237ClN3O4) + H]+,
33%}, 358.0 {[(C18H1235ClN3O4) + H]+, 100%}, 326 (2%).
(4-Chlorophenyl)[2-(phenylsulfonyl)pyridine-3-yl]metha-
none, 9. The general procedure described for the preparation
of 1 was used, except 4-chlorobenzaldehyde was used in place
of 2-chlorobenzaldehyde for the condensation reaction. These
conditions produced compound 9 as a tan solid in 82% yield
with 92.5% purity; mp 164-165 °C; (Found C, 60.28; H, 3.35;
N, 3.97; S, 9.00; Cl, 10.10. C18H12ClNO3S requires C, 60.42;
H, 3.38; N, 3.91; S, 8.96; Cl, 9.91); νmax/cm-1 (KBr) 3615, 1676,
1289, 1155; δH (400 MHz, CDCl3) 7.47 (2H, d, J 8.4),
7.50-7.56 (2H, m), 7.57 (1H, dd, J 7.6, 4.4), 7.59-7.66 (1H,
m), 7.73 (1H, dd, J 7.6, 1.6), 7.76 (2H, ddd, J 8.4, 2.4, 1.6),
7.98-8.04 (2H, m), 8.80 (1H, dd, J 4.4, 1.6); δC (100 MHz,
CDCl3) 126.4, 129.1, 129.2, 129.3, 131.3, 134.1 (6 × CH),
135.0, 135.5 (2 × C), 136.8 (CH), 138.5, 140.7 (2 × C), 150.8
(CH), 155.5, 191.9 (2 × C); HRMS (ES+): Exact mass
calculated for C18H13ClNO3S [M + H]+ 358.0305. Found
358.0301; m/z (ES+) 359.9 {[(C18H1237ClNO3S) + H+], 28%},
358.0 {[(C18H1235ClNO3S) + H+], 66%}.
2-{[2-(2-Chlorobenzoyl)phenyl]sulfonyl)pyridin-3-yl}(2-
chlorophenyl)methanone, 7. Tetrahydrofuran (40 mL) and
diisopropylamine (6.6 mL, 46.8 mmol, 2.55 equiv) were cooled
to -70 °C, and n-butyllithium (1.8 M, 22 mL, 39.6 mmol, 2.16
equiv) was added over 30 min at <-70 °C. The reaction mixture
was stirred for 30 min at -70 °C. A solution of sulfone 2 (4.02
g, 18.3 mmol, 1.0 equiv) was added in THF (12 mL) over 20
min, at <-70 °C. The reaction mixture was then stirred at -70
°C for 3.5 h. 2-Chlorobenzaldehyde (4.3 mL, 38.4 mmol, 2.1
equiv) was added over 15 min to the reaction mixture at <-70
°C. The reaction mixture was stirred for 1 h at -70 °C. Aqueous
3 N HCl (30 mL) was added over 30 min as the reaction mixture
was allowed to warm to ∼10 °C, and the layers were separated.
The aqueous layer was extracted with ethyl acetate (20 mL),
and the combined organic layers were stirred with saturated
aqueous sodium bicarbonate (20 mL) for 15 min. The layers
were separated, and the organic layer was washed with brine
(20 mL), dried with MgSO4, and concentrated under reduced
pressure to yield the alcohol (6.41 g) as an orange oil. By HPLC
analysis the mixture was found to contain 61.9% of the alcohol
5, 1.5% of 2, 31.6% 2-chlorobenzaldehyde, 4.4% of the bis-
impurity. Ethyl acetate (50 mL) was added followed by aqueous
KBr solution (10% w/w, 13.1 g, 0.6 equiv) and TEMPO (204
mg, 1.30 mmol, 0.07 equiv), and the reaction mixture was
cooled to 0 °C. A mixture of aqueous sodium hypochlorite (180
g of standard lab bleach, titrated as 2.5%, 3.5 equiv), sodium
bicarbonate (2.46 g, 29.1 mmol, 1.6 equiv) was added over 1 h
while allowing the reaction mixture to warm to room temper-
ature and stir for 17 h. The mixture was filtered to give the
bis-adduct 7 (0.59 g) as a white solid which was 94% pure by
HPLC, mp 209-212 °C; νmax/cm-1 (KBr) 3000, 1681, 1579,
1433, 1292; δH (300 MHz, CDCl3) 7.17-7.52 (10H, m), 7.65
(1H, ddd, appears as an overlapping td, J 7.5, 7.4, 1.3),
7.69-7.77 (1H, m), 7.82 (1H, dd, J 7.8, 1.6), 8.43 (1H, dd, J
7.8, 1.3), 8.58 (1H, dd, J 4.7, 1.6); δH (75.5 MHz, CDCl3) 125.7,
126.5, 126.6, 129.4, 130.7, 131.00, 131.01, 132.5, 132.6, 132.9,
133.23, 133.24, 133.3 (13 × CH), 133.5, 133.9, 134.8, 135.6,
135.8 (5 × C), 137.9 (CH), 138.0, 140.1 (2 × C), 149.7 (CH),
158.1, 191.3, 193.5 (3 × C); HRMS (ES+): Exact mass
calculated for C25H1635Cl2NO4S [(M + H)+], 496.0178. Found
496.0166, m/z (ES+) 500.0 {[(C25H1537Cl2NO4S) + H]+, 22%},
497.9{[(C25H1537Cl35ClNO4S)+H]+,80%},496.0{[(C25H1535Cl2-
NO4S) + H]+, 100%}.
Acknowledgment
We thank Mr. Michael Heller and Jeffery Lewis for the pilot-
plant synthesis of 1 and Otis Williams for providing engineering
support. We also thank Dr. S. E. Lawrence and C. E. Elcoate,
Department of Chemistry UCC, for project contributions. In
addition, Enterprise Ireland is acknowledged for financial
support of this project.
Supporting Information Available
1H and 13C NMR characterization data for key compounds
1, 5, 7, 8, and 9 are reported. This material is available free of
charge via the Internet at http://pubs.acs.org.
Received for review June 7, 2010.
OP100157Q
1238
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development