Acrylate as an efficient dimethylamine trap for the practical synthesis of 1-tert-butyl-4-piperidone via transamination

Article abstract of DOI:10.1021/op049860g

Efficient trapping of dimethylamine was the key to success in the transamination of 1,1-dimethyl-4-oxopiperidinium iodide with tert-butylamine to afford 1-tert-butylpiperidin-4-one in high yield. The use of sodium acrylate was found to provide an elegant way to both trap dimethylamine and provide a convenient method to allow purification in the subsequent extraction.

Full text of DOI:10.1021/op049860g

Organic Process Research & Development 2004, 8, 939−941
Communications to the Editor
Acrylate as an Efficient Dimethylamine Trap for the Practical Synthesis of
1-tert-Butyl-4-piperidone via Transamination
Joseph S. Amato,*,† John Y. L. Chung,*† Raymond J. Cvetovich,† Robert A. Reamer,† Dalian Zhao,† George Zhou,‡ and
Xiaoyi Gong‡
Department of Process Research and Department of Analytical Research, Merck Research Laboratories, P.O. Box 2000,
Rahway, New Jersey 07065, U.S.A.
Abstract:
desired piperidine is prepared by an exchange reaction
between 4-oxo-piperidinium iodide 5 and a primary amine
(Scheme 2). However, the major shortcoming of this ap-
proach is that exchange reactions frequently do not go to
completion due to unfavorable equilibria. Furthermore,
hindered amines tended to produce bis-amine Michael
adducts.
Efficient trapping of dimethylamine was the key to success in
the transamination of 1,1-dimethyl-4-oxopiperidinium iodide
with tert-butylamine to afford 1-tert-butylpiperidin-4-one in high
yield. The use of sodium acrylate was found to provide an
elegant way to both trap dimethylamine and provide a conve-
nient method to allow purification in the subsequent extraction.
Scheme 2
N-Substituted 4-piperidones are important synthetic build-
ing blocks to a considerable number of pharmacologically
active agents. This pharmacophore is most notably prevalent
in many CNS, antiallergic, and cardiovascular agents. In
connection with our drug development program, we required
kilogram quantities of 1-tert-butylpiperidin-4-one (1). The
classical synthesis of 1 in 14-26% yield reported by
Robinson1 and Fankhauser2 proceeded via a three-step
sequence involving a bis-Michael addition of tert-butylamine
with ethyl acrylate, followed by Dieckmann cyclization and
saponification-decarboxylation (Scheme 1). Because of the
steric hindrance of the tert-butyl group, the second Michael
addition to generate 3 proceeded in only a 27-34% yield
after heating at reflux for 4-14 days.
Utilizing a modified procedure using toluene/water,
K2CO3, and 10 equiv of tert-butylamine, the best assay yield
by NMR of 1 from 5 was ∼70% as limited by the equilibria,
which gave 40% isolated yield after distillation. Studies on
solvent effects revealed that DMSO was effective in enhanc-
ing the reaction rate and shifting the equilibrium in favor of
the desired product. Thus, the reaction in DMSO using excess
tert-butylamine with distillation to remove dimethylamine
gave a 95% assay yield and an 85% isolated yield on a 50
g scale, but on a 255 g scale the yield fell to 50%. We
attributed this to the decreased efficiency in the evaporative
removal of dimethylamine on scale-up due to reduced surface
area relative to volume leading to longer reaction times.
Addition of K2CO3 to the above reaction mixture afforded
near quantitative yields on small scales but highly variable
yields on larger scales. In these reactions run with K2CO3,
we consistently observed significant amounts of solid
deposited on the sides of the vessel above the reaction
mixture as the reaction proceeded. By NMR examination,
the solid was identified to be a dimethylamine derivative,
most likely dimethylammonium carbonate/carbamate. Thus
the formation of these materials physically removed dim-
ethylamine from the reaction, and the reaction was driven
forward to product. When the solid deposit was thrown back
into the reaction mixture, we observed an immediate decrease
of product 1 as monitored by real time FTIR analysis,
supporting that dimethylamine has been reintroduced and
Scheme 1
An alternate approach to N-substituted piperidones de-
veloped by Mistryukov3 and later modified by Kuehne4 and
others5 has some advantages over the classical conditions.
The troublesome bis-Michael addition is avoided, since the
* To whom correspondence should be addressed. E-mail: john_chung@
merck.com.
† Department of Process Research.
‡ Department of Analytical Research.
(1) Robinson, J. B.; Thomas, J. J. Chem. Soc. 1965, 2270-2271.
(2) Fankhauser, R.; Grob, C. A.; Krasnobajew, V. HelV. Chim. Acta 1966, 49,
690-695.
(3) Mistryukov, E. A.; Aronova, N. I.; Kucherove, V. F. IzV. Akad. Nauk SSSR,
Ser. Khim. 1961, 932-933.
(4) Kuehne, M. E.; Muth, R. S. J. Org. Chem. 1991, 56, 2701-2712.
(5) (a) Tschaen, D. M.; Abramson, L.; Cai, D.; Desmond, R.; Dolling, U.-H.;
Frey, L.; Karady, S.; Shi, Y.-J.; Verhoeven, T. R. J. Org. Chem. 1995, 60,
4324-4330. (b) Tortolani, D. R.; Poss, M. A. Org. Lett. 1999, 1, 1261-
1262. (c) Faul, M. M.; Kobierski, M. E.; Kopach, M. E. J. Org. Chem.
2003, 68, 5739-5741.
10.1021/op049860g CCC: $27.50 © 2004 American Chemical Society
Published on Web 10/12/2004
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
939

Scheme 3
Scheme 4
promoting the equilibria in the reverse direction. Further
studies on removal of dimethylamine by distillation and
displacement with tert-butylamine were only marginally
effective and required large amounts of tert-butylamine.
To gain further understanding, we studied the reaction
by NMR and FTIR. Upon mixing of all reagents, all of salt
5 was consumed within 5 min with little product 1 formation.
NMR analysis showed that the major product at this point
was compound 7 and that no olefinic substances were
present, as these must be highly reactive and short-lived.
Therefore the key for successful reaction would be to
selectively remove dimethylamine from the reaction. The
studies also showed that rapid removal is also critical for
high yields of 1, since it is not completely stable under the
reaction condition.
A number of solid-phase absorbents were screened for
the ability to absorb dimethylamine, and 13X molecular
sieves (8-12 mesh) were found to be the most effective.
Reaction reflux liquors were allowed to pass through a bed
of 13X sieves (2 mL sieves per gram of salt 5) and returned
to the reaction. The progress of the reaction was monitored
by online FTIR and offline GC, and when the reaction rate
slowed, the molecular sieves were replaced with fresh sieves.
Using this setup, treatment of salt 5 (up to 1 mol scale) with
excess tert-butylamine in DMSO and an equivalent of
Hunig’s base, the reaction was typically completed in 4-6
h with two changes of sieves to give an ∼95% assay yield
(Scheme 3). However, on multi-kilogram scale, the physical
dynamics of the system changed significantly due to
decreased surface area relative to volume, and extended
reaction time (>24 h) was required. The yield, as a
consequence, dropped to ∼75% due to formation of poly-
mers. These polymers had been previously observed in the
distillation of the product.
This led us to explore dimethylamine traps that would
react covalently. Acrylate esters, such as methyl- and tert-
butyl acrylates and methyl crotonate, proved to be highly
reactive traps for dimethylamine with good selectivity over
tert-butylamine in the ratios of 4-15:1 with the more
hindered acrylate achieving the better selectivity. In DMSO
at 60 °C, reactions were typically complete within 2-3 h
with a near quantitative yield of 1. To facilitate the isolation
of 1 and the removal of the acrylates and their amine adducts,
the reaction mixture was subjected to aqueous NaOH or TFA
to hydrolyze the methyl 3-alkylamino-propanoates or the tert-
butyl esters, respectively. Product 1 could then be isolated
by a simple acid-base extractive procedure, but significant
yield erosion (60% isolated yield) was observed due to
decomposition under the saponification conditions.
to observe the facile nature of this reaction. Furthermore,
water could be used in place of DMSO as the reaction
solvent. Thus heating a solution of salt 5 with excess tert-
butylamine in water in the presence of acrylic acid sodium
salt (from 5 eq acrylic acid and 4.6 eq of 10 M NaOH) at
65 °C for 1-2 h gave a 95% GC assay yield of N-tert-
butylpiperidone 1. The byproducts 3-(dimethylamino)-bu-
tanoate 11 and 3-tert-butylamino-butanoate 12 salts remained
in the water layer during the extraction of the product into
ethyl acetate. Evaporation of the organic extracts afforded
essentially pure 1 in 87% assay yield. This material was
assayed by GC to have a purity of 99.7 area % and 96.5 wt
% and was suitable for Ra-Ni reduction of the ketone
without catalyst poisoning problems. If required, the ketone
could be further purified by distillation under reduced
pressure to afford material of >99.9% purity. This process
was demonstrated on a half-mole scale with similar ef-
ficiency.7 The facile nature of this method should allow scale-
up without issues.
(7) Experimental Procedure (acrylate method): To a solution of acrylic acid
(176 g, 2.5 mol) in water (500 mL) was slowly added 10 N sodium
hydroxide (230 mL, 2.3 mol). Then 1,1-dimethyl-4-oxopiperidinium iodide
(5)10 (128 g, 0.5 mol) and tert-butylamine (1000 mL) were added, and the
resulting solution was heated at reflux (65 °C) for 90 min. The product
formation was monitored by GC assay which typically peaked at 90 min.
After completion of reaction the excess tert-butylamine was removed at 25
°C/40 Torr. The product is volatile, and thus excessive heating must be
avoided. The mixture was then extracted with ethyl acetate (500 mL), and
the aqueous was back-extracted twice with ethyl acetate (250 mL). The
combined organic was washed twice with aq NaCl (200 mL) and
concentrated in vacuo at 20 °C to give 70 g of 1 as an oil with a purity of
96.5 wt % and 99.7 area % by GC assay (87% assay yield). Analytically
pure sample (99.94 GC%) was obtained by distillation (∼109 °C/∼50 Torr
and ∼58 °C/∼1 Torr; lit.1 bp 92-94 °C/9 mm): 1H NMR (400 MHz,
DMSO-d6) δ 2.76 (t, J ) 6.1 Hz, 4 H), 2.29 (t, J ) 6.1 Hz, 4 H), 1.07 (s,
9 H); 13C NMR (100 MHz, DMSO-d6) δ 208.83, 53.48, 45.61, 41.67, 26.10.
The HCl salt of 1 was prepared from 2-propanol: 1H NMR (600.1 MHz,
DMSO-d6) δ 11.50 (br s, 1 H), 3.70-3.65 (m, 2 H), 3.44-3.37 (m, 2 H),
3.10 (ddd, J ) 16.6, 12.8, 6.0 Hz, 2 H), 2.45-2.40 (m, 2 H), 1.41 (s, 9 H);
13C NMR (150.9 MHz, DMSO-d6) δ 203.24, 62.92, 44.33, 37.29, 24.05.
(Distillation method): A mixture of 1,1-dimethyl-4-oxopiperidinium iodide
(5) (51 g, 200 mmol), tert-butylamine (200 mL), K2CO3 powder (30 g, 217
mol), and DMSO (200 mL) was heated at reflux (65 °C) under a condenser
(condenser temperature at 35-40 °C) for 4 h. Then the mixture was slowly
distilled over 1 h while fresh tert-butylamine (100 mL) was slowly added.
After cooling to 20 °C, added water (1 L) and extracted with MTBE (300
mL). The aqueous layer was back-extracted with MTBE (100 mL). The
combined organic was washed with aq NaCl and concentrated to an oil.
Distillation of the oil under reduced pressure (bp 110 °C/45 mm) afforded
27 g (86%) of 1.
(8) (a) Lyle, R. E.; Adel, R. E.; Lyle, G. G. J. Org. Chem. 1959, 24, 342-345.
(b) Van Luppen, J. J.; Lepoivre, J. A.; Dommisse, R. A.; Alderweireldt, F.
C. Org. Magn. Reson. 1979, 12, 399-405.
(9) A possible mechanism for achieving substantial rate improvement by acrylic
acid or their esters is the intermediacy of acrylate adducts of 7. Requater-
nization of the dimethyamine on 7 would be expected to facilitate elimination
and be kinetically favored over acrylate addition to the nitrogen bearing a
tert-butyl group.
(10) Compound 5 was readily prepared from commercially available 1-methyl-
4-piperidone and methyl iodide in acetone (15 mL/g) at 25-30 °C. After
stirring for 2 h, product precipitated and was isolated by filtration in 98%
yield.
To avoid this hydrolytic workup, we explored the use of
acrylic acid as the dimethylamine trap.6 We were gratified
(6) (a) Salov, V. N.; Zil’berman, E. N.; Krasnov, V. L. IzV. Vyssh. Uchebn.
ZzVed. Khim. Kh. Tekh. 1985, 28, 21-25. (b) Gros, C.; Chollet, H.; Mishra,
A. K.; Guilard, R. Synth. Commun. 1996, 26, 35-47. (c) Siaugue, J.-M.;
Segat-Dioury, F.; Sylvestre, I.; Favre-Reguillon, A.; Foos, J.; Madic, C.;
Guy, A. Tetrahedron 2001, 57, 4713-4718.
940
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development

Salts of 4-piperidones are known to form carbonyl
hydrates in aqueous solution.8 NMR studies of ketone 1 as
the free base and the monohydrochloride salt in DMSO-d6
showed that both species exist predominantly in the keto-
form (>98%) with less than 2% as the hydrate of the
carbonyl group even in the presence of residual water.
In conclusion, we have determined the key to successful
preparation of N-tert-butylpiperidone 1 via transamination
reaction of salt 5 was efficient removal of dimethylamine
and have evaluated several removal methods. Methods
involving physical means such as distillation and adsorption
by solid absorbents worked well for small scale preparations
but were not suitable for large scale. We discovered sodium
acrylate, which reacted with dimethylamine covalently in
water, to be the most efficient and practical trap for a scalable
synthesis. This procedure produced product 1 in high yield
and high purity, wherein purification by distillation was not
necessary. The accelerated rate of reaction also prevented
polymerization observed in other methods.9
Received for review July 30, 2004.
OP049860G
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
941