A new modified "montanari oxidation process" by means of chlorine dissolved in the reaction solvent as oxidant and TEMPO as catalyst: Oxidation of 3-S-quinuclidinol to 3-quinuclidinone

Article abstract of DOI:10.1021/op010096x

Results from a process improvement project for preparation of 3-R-quinuclidinol, a highly valuable intermediate for several muscarine-active compounds are described. The studied process was based on the kinetic resolution of racemic 3-quinuclidinol and th

Full text of DOI:10.1021/op010096x

Organic Process Research & Development 2002, 6, 197−200
A New Modified “ Montanari Oxidation Process” by Means of Chlorine Dissolved
in the Reaction Solvent as Oxidant and TEMPO as Catalyst: Oxidation of
3-S-Quinuclidinol to 3-Quinuclidinone
Hans-Rene´ Bjørsvik,*,† Lucia Liguori,† Francesca Costantino,‡,§ and Francesco Minisci‡
UniVersity of Bergen, Department of Chemistry, Alle´gaten 41, N-5007 Bergen, Norway, and
Politecnico di Milano, Dipartimento di Chimica, Via Mancinelli 7 I-20131 Milano, Italy
Abstract:
constitute 50% of the raw materials, (2) reducing the variable
costs, (3) increasing exploitation of the raw materials, (4)
improving the throughput of the process, and (5) general
process development and optimisation.
Results from a process improvement project for preparation
of 3-R-quinuclidinol, a highly valuable intermediate for several
muscarine-active compounds are described. The studied process
was based on the kinetic resolution of racemic 3-quinuclidinol
and thus involved a large side-stream production. Our outline
for process improvement is based on that this side stream can
be recycled by a two-step sequence of oxidation and reduction
processes. Findings concerning the conversion of 3-S-quinucli-
dinol into 3-quinuclidinone by using an improved oxidation
process based on TEMPO (2,2,6,6-tetramethylpiperidin-1-oxyl
radical) and molecular chlorine are discussed in detail. The
developed oxidation procedure affords nearly quantitatively
yield in the conversion of 3-S-quinuclidinol into 3-quinuclidi-
none.
Thus, in this context, we have outlined a process involving
the recycling of the side-stream of 3-S-quinuclidinol 2 in a
two step-sequence oxidation and reduction process. The first
step is constituted of the current BSY kinetic resolution
process of R- and S-3-quinuclidinol indicated as pathway
(a) in Scheme 1. The side stream of 3-S-quinuclidinol 2 is
according to our process sketch submitted for oxidation
following pathway (b) to give 3-quinuclidinone 3. In the
subsequent step, the ketone 3 can either undergo a simple
reduction, following pathway (c) or a stereoselective reduc-
tion following pathway (d), thus providing a racemic mixture
of R- and S-3-quinuclidinol and 3-R-quinuclidinol 1, respec-
tively. The net process constituted by the pathways (a), (b),
and (c) ideally converts all of the racemic material via the
achiral ketone 3 into the desired R form 1, and can be
considered as a dynamic kinetic resolution process. However,
an alternative to the sketched process is to follow the
pathways (e) and (d), where pathway (e) is exactly the same
oxidation procedure as applied in the pathway (b), while
pathway (d) requires a new efficient enantioselective reduc-
tion process that affords 3-R-quinuclidinol only, eventually
in a high enantiomeric excess.
Introduction
3-R-Quinuclidinol (R-1-azabicyclo[2.2.2]octan-3-ol) 1 is
used as a building block in the synthesis of muscarine M1
and M3 agonists as well as for muscarine M3 antagonists.
See for example the recent review by Broadley and Kelly1
concerning muscarinic receptor agonist and antagonist.
Talsaclidine fumarate (WAL 2014 FU),2-4 cevimeline HCl,5,6
and YM 9057 are all examples of products where 3-R-
quinuclidinol constitutes an integral part of the molecular
entity. These products have shown potential in the treatment
of Alzheimer’s disease, Sjo¨gren syndrome, and urinary
incontinence, respectively. 3-R-Quinuclidinol 1 constitutes
therefore a highly valuable building block for the syn-
thetic production of several important pharmaceutical chemi-
cals.
Methods and Results
The chemical literature reports several methods to be
generally applicable for oxidising secondary alcohols into
their corresponding ketones.8 A number of these were tested
for the oxidation step, pathway (b) of the outlined process
in Scheme 1. Nevertheless, most of the methods tried
afforded only low yields, did not operate at all, and ultimately
proved to be unsuitable for large-scale applications. In the
screening phase for oxidation methods and processes, organic
nitroxyl radicals also attracted our attention as potential
oxidants for our process, since such species are reported9,10
to be good oxidants for oxidising primary and secondary
alcohols.
3-R-Quinuclidinol 1 can be obtained by kinetic resolution
of racemic mixtures of R- and S-3-quinuclidinol. Borregaard
Synthesis, Norway (BSY), operates such a process, and it
was this that was the subject of a process improvement
project with focus on (1) decreasing the side streams, that
* To whom correspondence should be sent. E-mail: Hans.Bjorsvik@kj.uib.no.
† University of Bergen.
‡ Politecnico di Milano.
§ Present address: Industriale Chimica, Saronno (VA), Italy.
(1) Broadley, K. J.; Kelly, D. R. Molecules 2001, 6, 142-193.
(2) Drugs Future 1998, 24 (1), 79.
(8) Haines, A. H. Methods for the oxidation of organic compounds: alcohols,
alcohol deriVatiVes, alkyl halides, nitroalkanes, alkyl azides, carbonyl
compounds, hydroxyarenes and aminoarenes. Best synthetic methods;
Academic Press: London, 1988; pp 1-467.
(9) de Nooy, A. E. J.; Besemer, A. C.; van Bekkum, H. Synthesis 1996, 1153.
(10) Wicha, J.; Zarecki, A. Tetrahedron Lett. 1974, 35, 3059.
(3) Roma`n, G. Drugs Today 2000, 36 (9), 641.
(4) Drugs Future 2000, 26 (1), 61.
(5) Siggiqui, M F.; Levey, A I. Drugs Future 1999, 24 (4), 417.
(6) Sorbera, L. A.; Castaner, J. Drugs Future. 2000, 25 (6), 558
(7) Mealy, N.; Castaner, J. Drugs Future 1999, 24 (8), 871
10.1021/op010096x CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/14/2002
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
197

Scheme 1a
Table 1. Experimental results from oxidation experiments in
attempts to oxidise 1-aza-bicyclo [2.2.2] octan-3-ol to
1-aza-bicyclo [2.2.2] octan-3-onea
a (a) Kinetic resolution, (b) oxidation, (c) reduction, (d) stereoselective
reduction, (e) oxidation
oxidising system [O]
NaOCl/TEMPO
Ca(OCl)2/TEMPO
MCPBA/TEMPO
O2/CH3(CH2)2CHO/TEMPO
O2/t-Bu(OOCO)2/TEMPO
O2 (6 bar)/CuCl/TEMPO
C6H5CHO/Air (25 bar)/TEMPO
H2O2/Br2
solvents
yield of 3 (%) (gc)
1
2
3
4
5
6
7
8
CH2Cl2/H2O
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH3CN
EtOAc
16
10
78
traces
2
traces
5
traces
Scheme 2
EtOAc
a i) The reactions were carried out at 25 °C in 100 mL scale starting with
1.27 g (10 mmol) of the substrate.
Table 2. Experimental results from oxidation experiments in
attempts to oxidise 1-aza-bicyclo [2.2.2]octan-3-ol to
1-aza-bicyclo[2.2.2]octan-3-one using TEMPO and chlorinea
A particularly convenient method, involving a two-phase
system constituted by methylene chloride and water, using
hypochlorite as oxidant and 2,2,6,6-tetramethyl-piperidin-
1-oxyl (TEMPO) 5 as catalyst, was reported by Montanari
and co-workers.11-13 The mechanism for this oxidation
reaction is given in Scheme 2. It is worth noting that the
real oxidant for the alcohol in this cycle is the oxo-
piperidinium compound 5a (2,2,6,6-tetramethyl-1-oxo-pip-
eridinium). While hypochlorite is indicated as the oxidant
for 2,2,6,6-tetramethyl-piperidin-1-ol 5c and TEMPO 5,
several other oxidants may also be used.
amounts (%) (gc)
time for bubbling
Cl2 (min)
reaction
time
w/w 5
3
2
6
1
2
3
10%
5%
1%
30
35
60
1 h 45 min 88.0 <1
9.2
2 h 40 min 73.8 <1 26.1
2 h 60.0 <1 37.5
a i) The reactions were carried out at a temperature of 25 °C in 100 mL
methylene chloride using 1.27 g (10 mmol) of the alcohol 2.
atom economy, as equivalents of m-chlorbenzoic acid are
produced as side product.
However, our trials using the “Montanari procedure”
revealed rather unsatisfactory results when utilised for the
alcohol 1: conversion and selectivity where low. In our
opinion this is due to the presence of the basic nitrogen,
which strongly increases the solubility of the alcohol 1 in
the aqueous phase. Other oxidative systems used in combi-
nation with TEMPO 5, such as molecular oxygen and
aldehydes (acetaldehyde, benzaldehyde) in combination with
Cu salts, and peracids generated in situ from acetic anhydride
and hydrogenperoxide, gave only trivial results. We believe
that all of these poor results were due to the presence of the
basic nitrogen in the alcohol 2. However, when m-chloro-
perbenzoic acid (MCPBA) was used as oxidant for 5c and 5
(entry 3 Table 1) a quite satisfactory result was achieved.
This procedure has however the disadvantages of defective
We thus investigated a modification of the Montanari
procedure11-13 developed by Yamaguchi et al.14 involving
TEMPO 5, using anhydrous sodium carbonate in methylene
chloride through which is passed a stream of chlorine gas.
We have found that this procedure works moderately
satisfactory with the alcohol 1. In these experiments,
relatively high yields of the ketone 3 were achieved even
without rigorous optimisation of the oxidation process (Table
2). A chlorinated product was also formed during the
reaction. GC-MS analysis suggested a monochlorinated
derivative of the ketone 3, most probably formed due to a
subsequent electrophilic chlorination of the formed ketone
3 to give 2-chloro-1-azabicyclo[2.2.2]octan-3-one 6. In
addition to the formation of the side product, several other
concerns were also present for the use of such an oxidation
process for large-scale process, namely, the handling of
chlorine gas during the oxidation process and the relative
(11) Anelli, P. L.; Biffi, C.; Montanari, F. J. Org. Chem. 1987, 52, 2559.
(12) Anelli, P. L.; Banfi, S.; Montanari, F.; Quici, S. J. Org. Chem. 1989, 54,
2970.
(13) Anelli, P. L.; Banfi, S.; Montanari, F.; Quici, S. Org. Synth. 1990, 69,
212.
(14) Yamaguchi, M.; Takata, T.; Endo, T. Bull. Chem. Soc. Jpn. 1990, 63, 947.
198
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

Table 3. Experimental results from oxidation experiments in
attempts to oxidise 1-aza-bicyclo[2.2.2] octan-3-ol to
residence time with possibility of higher througput in plant
scale.
1-aza-bicyclo[2.2.2]octan-3-one using TEMPO and chlorinea
Interestingly, when a reduced quantity of TEMPO 5 is
used under otherwise similar condition, the ketone 3 is in a
subsequent process chlorinated to give the chloroketone 6
in quite high yields. Especially when the chlorine gas is
bubbled through the reaction mixture, high yields are
afforded (Table 2). Further optimising by using response
surface methodology15 of this process may provide a method
for preparation for exactly this compound.
amount of Cl2
amounts (%) (gc)
1stAdd
chlorine
2ndAdd
chlorine
w/w 5 (%) t [min]
n
n
3
2
6
Experimental Section
Reaction Procedure Bubbling Cl2 through the Oxida-
tion Mixture. 1-Azabicyclo[2.2.2]octan-3-ol (10 mmol, 1.27
g), sodium carbonate (70 mmol, 7.4 g), and 2,2,6,6-
tetramethylpiperidin-1-oxyl (1-10% w/w) was mixed in
CH2Cl2 (100 mL). Chlorine gas was then bubbled through
the reaction mixture, approximately one bubble per each 2
s through a Pasteur pipet for a reaction time of 105-160
min.
Reaction Procedure Using Cl2 Dissolved in CH2Cl2.
The solution of dissolved chlorine in methylene chloride was
prepared by bubbling chlorine gas through CH2Cl2 (50
mL) for 30 min. The concentration of absorbed chlorine gas
was determined by adding KI (200 mg) in concentrated
CH3COOH (5 mL) and water (5 dr) to a sample of 10 mL
of the CH2Cl2/Cl2 solution. This solution was then titrated
with Na2S2O3 (0.1 N; 59 mL were used). The chlorine content
in CH2Cl2 was determined to be 1.18 M.
1
2
3
4
10
5
5
30
30
30
30
11.8
11.8
11.8
11.8
11.8
-
17.7
18.0
∼99
69
<1
28
<1
<1
-
-
-
∼99
1
78
15
a i) At the start of the reaction 11.8 mmol of chlorine (nchlorine
the reaction mixture. A second dropwise addition (n2ndAdd
chlorine) of chlorine to the
reaction mixture was carried out over a period of 30 min.
1stAdd ) was added to
high quantity of TEMPO 5 (10%) that was required. The
experiments utilising bubbling chlorine gas as oxidant with
TEMPO 5 as catalyst, revealed, however, that an oxidation
system based on using the oxo-piperidinium compound 5a
as the oxidant for the alcohol 2 is a promising direction in
which to develop the process.
In this context, our idea was to utilize the same solvent
as carrier for the chlorine gas as for preparing the solution
of the alcohol 2, TEMPO 5, and anhydrous sodium carbonate
to which the gas solution would be added. Thus, a solution
(∼1.2 M) of molecular chlorine in methylene chloride was
prepared and used in the series of experiments reported in
Table 3.
Entry 1 of Table 3 shows the result when 10% w/w of
TEMPO 5, the alcohol 2 and 2.36 equiv of chlorine are used.
This experiment affords quantitative transformation of the
alcohol 2, into desired product, the ketone 3, without forma-
tion of chloroketone 6. When the quantity of TEMPO 5 is
reduced to 5% w/w, a somewhat higher amount of chlorine
was necessary (totally 2.95 equiv) to achieve a quantitatively
transformation of the alcohol 2 (entry 3 of Table 3). Without
the following-up dropwise addition of chlorine, a partial
conversion of the alcohol is achieved (entry 2 of Table 3).
Entry 4 of Table 3 gives the result from a experiment using
only 1% w/w of TEMPO. In this experiment there was also
formed the chlorinated side product 6 in the same way as
for the experiment where chlorine gas was bubbled through
the reaction mixture; at higher concentrations of TEMPO 5
(entries 1 and 3) no detectable quantities of 6 (2-chloro-1-
azabicyclo[2.2.2]octan-3-one) were observed.
The oxidation reaction was carried out by dissolving
3-quinuclidinol (1.27 g, 10 mmol) in CH2Cl2 (10 mL) to
which anhydrous Na2CO3 (7.40 g, 70 mmol) and TEMPO
(1-10% w/w) were added. The CH2Cl2/Cl2 solution (0.0118
mol, 10 mL of 1.18 M) was added in one portion followed
by another (nchlorine
2ndAdd) mol of chlorine (see Table 3) dropwise
over a period of 30 min.
Work-up and Purification. The white solid was filtered
off, and the solvent was evaporated. Solvent system used
for TLC: CH2Cl2:MeOH/1:1. The ketone was isolated by
flash chromatography (CH2Cl2:MeOH/8:2) and characterised
1
by H and 13C NMR.
1H NMR (MeOD): 3.38-2.82 δ (m, 5H), 2.46-1.62 δ
(m, 6H).
13C NMR (MeOD): δ 219.4 (CdO), 57.64 (N-C-CO),
40.56 (CH2-N), 35.96 (C-CO), 27.62 (CH2).
Crystallisation Procedure 1. The red crude product (1.3
gram) was dissolved in a solvent mixture (1:1) of CH2Cl2
and CH3OH (10 mL). However, the crude product was not
completely soluble due to the presence of tars. The obtained
solution (orange-coloured) was filtered using a glass sinter
filter to remove the tars. Then, the solvent was removed
under vacuum to obtain about 150 mg of orange solid, which
was crystallised using 2-propanol (5 mL). The crystallised
product appeared as a white powder (isolated yield about
100 mg). The product was analysed on GC-MS, which
showed only the desired product.
Conclusions
Our modification of the Montanari procedure11-13 using
a saturated solution of Cl2 in methylene chloride shows
important advantages for large-scale production: (1) it avoids
the handling of chlorine as gas during the oxidation process,
(2) it affords excellent yields even with a reduced quantity
of catalyst, (3) the reaction rate is also increased using the
modified procedure, with the consequence of a shorter
(15) Box, G. E. P.; Draper, N. R. Empirical Model-Building and Response
Surfaces; Wiley: New York 1987, 1-669.
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
199

Crystallisation Procedure 2. The red crude product (760
mg) was dissolved in 2-propanol (15 mL). Charcoal (200
mg) was added to this solution, and the mixture was stirred
for 1.5 h at 100 °C. The solution was then filtered on a glass
sinter filter. A brown solid was precipitated from the filtrate.
Hence, a second filtration was performed to obtain a pale-
orange solution. The solvent was then removed under
vacuum. The residue was recrystallised by dissolving the
solid in ethanol (2 mL). Petroleum ether bp 30-60 °C (1
mL) was added to this solution to lower the solubility of the
ketone and hence initiate crystallisation. Isolated yield of the
ketone was 170 mg. The product was analysed on GC-MS,
which showed only the desired product.
It should be noted that experience indicates that an
immediate work-up with recrystallisation after oxidation
gives a higher isolated yield as well as a simpler work-up.
Acknowledgment
Borregaard Synthesis (Norway) is gratefully acknowl-
edged for financial support and the permission to publish
the present work. Professor George Francis is acknowledged
for discussions and linguistic support during the preparation
of this paper.
Received for review October 11, 2001.
OP010096X
200
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development