borane-amine adducts, which is a reflection of how tightly
the borane is complexed to the amine, does limit their
reactivity. Brown has shown that imines can be reduced with
various reactive tert-butyldialkylamine-borane adducts.17
Others have shown the same transformation can be ac-
complished using less reactive complexes such as dimethyl-
amine-borane and pyridine-borane if the reduction is
performed in the presence of an acid, which activates the
borane-amine complex.24,25
Figure 1. Proposed compounds detected with LC-MS during
NaBH4 reduction of 5.
For the present reduction, borane-tert-butylamine (BTBA)
and methanesulfonic acid in methylene chloride were used
as the reduction medium. BTBA was selected because the
tert-butylamine (bp 45 °C) liberated from the reducing agent
could be removed during a downstream distillation and would
not interfere with the isolation of 1 as the hydrochloride salt.
Methanesulfonic acid was used as an activator because it is
anhydrous and it formed a soluble complex with imine 5.
The initial reaction conditions involved the addition of 2
equiv of methanesulfonic acid to an equimolar mixture of
imine 5 and BTBA in methylene chloride. Acid (2 equiv)
was used to protonate both the liberated tert-butylamine and
the generated amine 3, thereby preventing the product from
complexing with the borane. The reduction was clean and
fast when these conditions were used; however there was a
significant exotherm, presumably due to multiple simulta-
neous exothermic reactions, so the order of addition was
changed. Addition of methanesulfonic acid (2 equiv) to imine
5 in methylene chloride still caused an appreciable temper-
ature rise, but examination with a reaction calorimeter (RC-
1) revealed that the exotherm was immediate and propor-
tional to the dose of methanesulfonic acid. The heat output
was significantly attenuated during the addition of the second
equivalent of acid (Figure 2) and may be explained by
considering that an acid-base reaction has occurred during
the addition of the first equivalent of acid only. The heat
generated during the addition of the second equivalent of
acid is due to dilution of the methanesulfonic acid. For the
total addition, the heat output was 77 kJ/mol of 5, which
corresponded to an adiabatic temperature rise (∆Tad) of 23
°C. Extended addition times were well tolerated, and the
combination of 5 and methansulfonic acid was stable in
solution for at least 16 h as long as moisture was excluded.
Addition of the mixture of 5 and methanesulfonic acid to
a suspension of BTBA in methylene chloride was also
significantly exothermic (Figure 3), but once again the
exotherm was immediate and directly related to the dose of
the imine-acid solution. The profile of the heat output
indicated that the reduction was quite fast, since heat
evolution ceased when the addition was finished.
Although the direct alkylation route was safer, since
reductive conditions were avoided, the projected costs of
goods would be high. A large excess of methylamine and
another relatively expensive starting material12 were required,
and a significant amount of processing was necessary to
provide product of acceptable purity.13 On the other hand,
the potential for reduction of the nitro group with sodium
borohydride prohibited scale-up of the reductive amination
route. The key to developing a scalable process was in
finding a reducing agent capable of selectively reducing the
imine functional group in the presence of the aryl nitro group.
Imine 5 could be conveniently formed in yields approach-
ing 90% by adding aqueous methylamine (2.5 equiv) to
3-nitrobenzaldehyde in heptane. The reaction was complete
within 30 min, and the product was isolated by filtering the
reaction mixture.14 Alternatively, 5 could be prepared from
the aldehyde by generating methylamine in situ from
methylamine hydrochloride and sodium hydroxide in a
mixture of methylene chloride and water. When the con-
densation was complete, the organic layer was separated and
the solvent was exchanged to hexanes, which caused
precipitation of imine 5 in ca. 90% yield.
Diborane and specific borane complexes have displayed
selectivity towards the reduction of many functional groups
in the presence of a nitro group.15-17 Of the borane complexes
commercially available, borane-amine complexes18,19 are
particularly attractive reagents for large-scale work because
many are air-stable solids or liquids.20-23 The stability of
(12) 3-Nitrobenzyl chloride was only available with a lead time of 12-16 weeks
and a cost of ca. $750-$1650/kg.
(13) Mixtures of the hydrochloride salts of 1 and 3 can be separated in a rather
labor-intensive procedure that involves dissolving the mixture in water and
extracting the dimer hydrochloride salt into methylene chloride. The aqueous
solution containing 1 is adjusted to pH 10 and extracted with methylene
chloride, and the solvent is exchanged with 2-propanol or ethanol. Dry
HCl is added to precipitate the salt, and heptane is added as an antisolvent.
Hydrochloride salt of 1 is then filtered and dried.
(14) The condensation is also successful under biphasic conditions using
methylamine hydrochloride, sodium hydroxide, and water. A higher charge
of water is used in this case to ensure the sodium chloride is extracted into
the aqueous layer and does not interfere with the isolation of the imine.
These conditions have not been demonstrated on scales larger than 5 g.
(15) Brown, H. C.; Subba Rao, B. C. J. Am. Chem. Soc. 1960, 82, 681-686.
(16) Yoon, N. M.; Pak, C. S.; Brown, H. C.; Krishnamurthy, S.; Stocky, T. P.
J. Org. Chem. 1973, 38, 2786-2792.
(17) Brown, H. C.; Kanth, J. V. B.; Dalvi, P. V.; Zaidlewicz, M. J. Org. Chem.
1999, 64, 6263-6274.
(18) Hutchins, R. O.; Learn, K.; Nazer, B.; Pytlewski, D.; Pelter, A. Org. Prep.
Proced. Int. 1984, 16, 335-372.
(19) Lane, C. F. Aldrichimica Acta 1973, 6, 51-58.
(20) Couturier, M.; Andresen, B. M.; Tucker, J. L.; Dube, P.; Brenek, S. J.;
Negri, J. T. Tetrahedron Lett. 2001, 42, 2763-2766.
(21) Couturier, M.; Tucker, J. L.; Andresen, B. M.; Dube, P.; Brenek, S. J.;
Negri, J. T. Tetrahedron Lett. 2001, 42, 2285-2288.
When the reduction was complete, the reaction mixture
was quenched with dilute aqueous ammonia. Addition of
aqueous ammonia to the reaction mixture did not lead to
addition control over the heat evolution or off-gas profile
(Figure 4). Although the aqueous ammonia was charged over
30 min, the exotherm and off-gassing occurred within the
first 5 min. The 400 W exothermic spike was approximately
(22) Couturier, M.; Tucker, J. L.; Andresen, B. M.; Dube, P.; Negri, J. T. Org.
Lett. 2001, 3, 465-467.
(23) Andrews, G. C.; Crawford, T. C. Tetrahedron Lett. 1990, 21, 693-696.
(24) Billman, J. H.; McDowell, J. W. J. Org. Chem. 1961, 26, 1437-1440.
(25) Pelter, A.; Rosser, R. M.; Mills, S. J. Chem. Soc., Perkin Trans. 1 1984,
717-720.
838
•
Vol. 9, No. 6, 2005 / Organic Process Research & Development