Reductive amination of aldehydes and ketones with sodium triacetoxyborohydride. Studies on direct and indirect reductive amination procedures

Article abstract of DOI:10.1021/jo960057x

Sodium triacetoxyborohydride is presented as a general reducing agent for the reductive amination of aldehydes and ketones. Procedures for using this mild and selective reagent have been developed for a wide variety of substrates. The scope of the reaction includes aliphatic acyclic and cyclic ketones, aliphatic and aromatic aldehydes, and primary and secondary amines including a variety of weakly basic and nonbasic amines. Limitations include reactions with aromatic and unsaturated ketones and some sterically hindered ketones and amines. 1,2-Dichloroethane (DCE) is the preferred reaction solvent, but reactions can also be carried out in tetrahydrofuran (THF) and occasionally in acetonitrile. Acetic acid may be used as catalyst with ketone reactions, but it is generally not needed with aldehydes. The procedure is carried out effectively in the presence of acid sensitive functional groups such as acetals and ketals; it can also be carried out in the presence of reducible functional groups such as C-C multiple bonds and cyano and nitro groups. Reactions are generally faster in DCE than in THF, and in both solvents, reactions are faster in the presence of AcOH. In comparison with other reductive amination procedures such as NaBH₃CN/MeOH, borane-pyridine, and catalytic hydrogenation, NaBH(OAc)₃ gave consistently higher yields and fewer side products. In the reductive amination of some aldehydes with primary amines where dialkylation is a problem we adopted a stepwise procedure involving imine formation in MeOH followed by reduction with NaBH₄.

Full text of DOI:10.1021/jo960057x

J . Org. Chem. 1996, 61, 3849-3862
3849
Red u ctive Am in a tion of Ald eh yd es a n d Keton es w ith Sod iu m
Tr ia cetoxybor oh yd r id e. Stu d ies on Dir ect a n d In d ir ect Red u ctive
Am in a tion P r oced u r es1
Ahmed F. Abdel-Magid,* Kenneth G. Carson, Bruce D. Harris, Cynthia A. Maryanoff, and
Rekha D. Shah
The R. W. J ohnson Pharmaceutical Research Institute, Department of Chemical Development,
Spring House, Pennsylvania 19477
Received J anuary 8, 1996X
Sodium triacetoxyborohydride is presented as a general reducing agent for the reductive amination
of aldehydes and ketones. Procedures for using this mild and selective reagent have been developed
for a wide variety of substrates. The scope of the reaction includes aliphatic acyclic and cyclic
ketones, aliphatic and aromatic aldehydes, and primary and secondary amines including a variety
of weakly basic and nonbasic amines. Limitations include reactions with aromatic and unsaturated
ketones and some sterically hindered ketones and amines. 1,2-Dichloroethane (DCE) is the preferred
reaction solvent, but reactions can also be carried out in tetrahydrofuran (THF) and occasionally
in acetonitrile. Acetic acid may be used as catalyst with ketone reactions, but it is generally not
needed with aldehydes. The procedure is carried out effectively in the presence of acid sensitive
functional groups such as acetals and ketals; it can also be carried out in the presence of reducible
functional groups such as C-C multiple bonds and cyano and nitro groups. Reactions are generally
faster in DCE than in THF, and in both solvents, reactions are faster in the presence of AcOH. In
comparison with other reductive amination procedures such as NaBH3CN/MeOH, borane-pyridine,
and catalytic hydrogenation, NaBH(OAc)3 gave consistently higher yields and fewer side products.
In the reductive amination of some aldehydes with primary amines where dialkylation is a problem
we adopted a stepwise procedure involving imine formation in MeOH followed by reduction with
NaBH4.
In tr od u ction
amine are mixed with the proper reducing agent without
prior formation of the intermediate imine or iminium
salt. A stepwise or indirect reaction involves the prefor-
mation of the intermediate imine followed by reduction
in a separate step.
The reactions of aldehydes or ketones with ammonia,
primary amines, or secondary amines in the presence of
reducing agents to give primary, secondary, or tertiary
amines, respectively, known as reductive aminations (of
the carbonyl compounds) or reductive alkylations (of the
amines) are among the most useful and important tools
in the synthesis of different kinds of amines. The
reaction involves the initial formation of the intermediate
carbinol amine 3 (Scheme 1) which dehydrates to form
an imine. Under the reaction conditions, which are
usually weakly acidic to neutral, the imine is protonated
to form an iminium ion 4.2 Subsequent reduction of this
iminium ion produces the alkylated amine product 5.
However, there are some reports that provide evidence
suggesting a direct reduction of the carbinol amine 3 as
a possible pathway leading to 5.3 The choice of the
reducing agent is very critical to the success of the
reaction, since the reducing agent must reduce imines
(or iminium ions) selectively over aldehydes or ketones
under the reaction conditions.
The two most commonly used direct reductive amina-
tion methods differ in the nature of the reducing agent.
The first method is catalytic hydrogenation with plati-
num, palladium, or nickel catalysts.2a,4 This is an
economical and effective reductive amination method,
particularly in large scale reactions. However, the reac-
tion may give a mixture of products and low yields
depending on the molar ratio and the structure of the
reactants.5 Hydrogenation has limited use with com-
pounds containing carbon-carbon multiple bonds and in
the presence of reducible functional groups such as
nitro6,7 and cyano7 groups. The catalyst may be inhibited
by compounds containing divalent sulfur.8 The second
method utilizes hydride reducing agents particularly
sodium cyanoborohydride (NaBH3CN) for reduction.9 The
successful use of sodium cyanoborohydride is due to its
stability in relatively strong acid solutions (∼pH 3), its
solubility in hydroxylic solvents such as methanol, and
its different selectivities at different pH values.10 At pH
The reductive amination reaction is described as a
direct reaction when the carbonyl compound and the
X Abstract published in Advance ACS Abstracts, May 1, 1996.
(1) Presented in part at the 33rd ACS National Organic Symposium,
Bozeman, Mo, J une 1993, Paper A-4. Preliminary communications: (a)
Abdel-Magid, A. F.; Maryanoff, C. A.; Carson, K. G. Tetrahedron Lett.
1990, 31, 5595. (b) Abdel-Magid, A. F.; Maryanoff, C. A. Synlett 1990,
537.
(2) The formation of imines or iminium ions was reported as possible
intermediates in reductive amination reactions in catalytic hydrogena-
tion methods, see (a) Emerson, W. S. Org. React. 1948, 4, 174 and
references therein. It was also proposed in hydride methods, see (b)
Schellenberg, K. A. J . Org. Chem. 1963, 28, 3259.
(4) (a) Emerson, W. S.; Uraneck, C. A. J . Am. Chem. Soc. 1941, 63,
749. (b) J ohnson, H. E.; Crosby, D. G. J . Org. Chem. 1962, 27, 2205.
(c) Klyuev, M. V.; Khidekel, M. L. Russ. Chem. Rev. 1980, 49, 14.
(5) Skita, A.; Keil, F. Chem. Ber. 1928, 61B, 1452.
(6) Roe, A.; Montgomery, J . A. J . Am. Chem. Soc. 1953, 75, 910.
(7) Rylander, P. N. In Catalytic Hydrogenation over Platinum
Metals; Academic Press, New York, 1967; p 128.
(8) Rylander, P. N. In Catalytic Hydrogenation over Platinum
Metals; Academic Press, New York, 1967; p 21.
(9) For a recent review on reduction of CdN compounds with hydride
reagents see: Hutchins, R. O., Hutchins, M. K. Reduction of CdN to
CHNH by Metal Hydrides. In Comprehensive Organic Synthesis; Trost,
B. N., Fleming, I., Eds.; Pergamon Press: New York, 1991; Vol. 8.
(3) Tadanier, J .; Hallas, R.; Martin, J . R.; Stanaszek, R. S. Tetra-
hedron 1981, 37, 1309
S0022-3263(96)00057-6 CCC: $12.00 © 1996 American Chemical Society

3850 J . Org. Chem., Vol. 61, No. 11, 1996
Abdel-Magid et al.
Sch em e 1
3-4 it reduces aldehydes and ketones effectively, but this
reduction becomes very slow at higher pH values.11 At
pH 6-8, the more basic imines are protonated preferen-
tially and reduced faster than aldehydes or ketones.10
This selectivity allows for a convenient direct reductive
amination procedure. The literature is replete with
publications that document the use of sodium cyanoboro-
hydride in reductive amination reactions.12 Limitations
are that the reaction may require up to a fivefold excess
of the amine,10 is usually slow and sluggish with aromatic
ketones10 and with weakly basic amines,13 and may result
in the contamination of the product with cyanide.14 The
reagent is highly toxic15 and produces toxic byproducts
such as HCN and NaCN upon workup.
Other reported reductive amination reagents include
borane-pyridine,13a Ti(OiPr)4/NaBH3CN,13b borohydride
exchange resin,16a Zn/AcOH,16b NaBH4/Mg(ClO4)2,16c and
Zn(BH4)2/ZnCl2.16d In addition, there are some reports
of electrochemical reductive amination reactions.17
In our work on hydride-induced reductive aminations
of aldehydes and ketones, we sought an alternative to
the toxic sodium cyanoborohydride to eliminate the risk
of residual cyanide in the product and in the workup
waste stream, particularly for large scale reactions. Af-
ter surveying many of the commercially available hy-
dride reagents, we selected sodium triacetoxyborohydride
[NaBH(OAc)3].18 This borohydride reagent is mild and
exhibits remarkable selectivity as a reducing agent. It
reduces aldehydes selectively over ketones,18 except for
â-hydroxy ketones which can be reduced selectively to
give 1,3-trans diols.19 The steric and the electron-
withdrawing effects of the three acetoxy groups stabilize
the boron-hydrogen bond and are responsible for its mild
reducing properties.20 Our selection was also based on
the results of reductive alkylation of amines using sodium
borohydride in neat liquid carboxylic acids reported
earlier by Gribble et al.21
In this paper we report the results of our comprehen-
sive investigation of the scope and limitations of sodium
triacetoxyborohydride in a procedure for direct reductive
amination of aldehydes and ketones with a variety of
aliphatic and aromatic amines. This report also includes
an alternative stepwise route for the reductive amination
of aldehydes with primary amines which involves the
preformation of imines and their subsequent reduction
with NaBH4 in one-pot reactions.
Resu lts a n d Discu ssion s
The direct reductive amination reactions were carried
out in 1,2-dichloroethane (DCE), tetrahydrofuran (THF),
or acetonitrile. The standard reaction conditions are as
follows: a mixture of the carbonyl compound and the
amine (0-5% molar excess) in the desired solvent is
stirred with 1.3-1.6 equiv of sodium triacetoxyborohy-
dride under a nitrogen atmosphere at room temperature.
In some reactions, acetic acid (1-2 mol equiv) is added
to the mixture. The progress of the reaction is followed
by GC and GC/MS analysis. The results from various
reductive amination reactions of ketones and aldehydes
are listed in Tables 1 and 2, respectively. Solvents such
as water or methanol are not recommended. Reactions
in methanol resulted in a fast reduction of the carbonyl
compound, and the reagent decomposed in water.
(10) Borch, R. F.; Bernstein, M. D.; Durst, H. D. J . Am. Chem. Soc.
1971, 93, 2897.
(11) Borch, R. F., Durst, H. D. J . Am. Chem. Soc. 1969, 91, 3996.
(12) (a) Hutchins, R. O.; Natale, N. R. Org. Prep. Proced. Int. 1979,
11(5), 201. (b) Lane, C. F. Synthesis 1975, 135.
(13) Occasional use of weakly basic or nonbasic amines was reported,
see for example: (a) Pelter, A., Rosser, R. M., Mills, S. J . Chem. Soc.,
Perkin Trans. 1 1984, 717. (b) Mattson, R. J ., Pham, K. M.; Leuck, D.
J .; Cowen, K. A. J . Org. Chem. 1990, 55, 2552. (c) Borch, R. F.; Hassid,
A. I. J . Org. Chem. 1972, 37, 1673. (d) Marchini, P.; Liso, G.; Reho, A.;
Liberatore, F.; Moracci, F. M. J . Org. Chem. 1975, 40, 3453.
(14) (a) The product from large scale reduction of the imine (i) with
sodium cyanoborohydride was contaminated with cyanide. (b) A similar
result was reported recently: Moormann, A. E. Synth. Commun. 1993,
23, 789.
(a ) Red u ctive Am in a tion of Keton es. The results
in Table 1 show that the reductive amination of a wide
variety of cyclic and acyclic ketones with primary and
(18) (a) Gribble, G. W.; Ferguson, D. C. J . Chem. Soc., Chem.
Commun. 1975, 535. (b) Nutaitis, C. F.; Gribble, G. W. Tetrahedron
Lett. 1983, 24, 4287. (c) Gribble, G. W. In Encyclopedia of Reagents
for Organic Synthesis; Paquette, L. A., Ed., J ohn Wiley and Sons: New
York, 1995; Vol. 7, p 4649.
(19) See for example: (a) Saksena, A. K.; Mangiaracina, P. Tetra-
hedron Lett. 1983, 24, 273. (b) Evans, D. A.; Chapman, K. T.
Tetrahedron Lett. 1986, 27, 5939. (c) Evans, D. A., Chapman, K. T.;
Carreira, E. M. J . Am. Chem. Soc. 1988, 110, 3560.
(15) For information on the safety data and health hazards associ-
ated with sodium cyanoborohydride see: The Sigma-Aldrich Library
of Chemical Safety Data, 1st ed.; Lenga, R. E., Ed., Sigma-Aldrich
Corp.: Milwaukee, 1985, p 1609.
(16) (a) Yoon, N. M.; Kim, E. G.; Son, H. S.; Choi, J . Synth. Commun.
1993, 23, 1595. (b) Micovic, I. V.; Ivanovic, M. D.; Piatak, D. M.; Bojic,
V. Dj. Synthesis 1991, 1043. (c) Brussee, J .; van Benthem, R. A. T. M.;
Kruse, C. G.; van der Gen, A. Tetrahedron: Asymmetry 1990, 1, 163.
(d) Bhattacharyya, S.; Chatterjee, A.; Duttachowdhhury, S. K. J . Chem.
Soc., Perkin Trans. 1 1994, 1.
(17) (a) Pienemann, T.; Schafer, H.-J . Synthesis 1987, 1005. (b)
Smirnov, Yu. D.; Tomilov, A. P. J . Org. Chem. U.S.S.R. 1992, 28(1),
42. (c) Smirnov, Yu. D.; Pavlichenko, V. F.; Tomilov, A. P. J . Org. Chem.
U.S.S.R. 1992, 28(3), 374.
(20) Gribble, G. W.; Nutaitis, C. F. Org. Prep. Proced. Int. 1985, 17,
317.
(21) Earlier work by Gribble et al. demonstrated the potential of
triacyloxyborohydrides generated from NaBH4 in neat liquid carboxylic
acids in reductive alkylation of amines: (a) Gribble, G. W.; Lord, P.
D.; Skotnicki, J .; Dietz, S. E.; Eaton, J . T.; J ohnson, J . L. J . Am. Chem.
Soc. 1974, 96, 7812. (b) Gribble, G. W.; J asinski, J . M.; Pellicone, J .
T.; Panetta, J . A. Synthesis 1978, 766.

Reductive Amination of Aldehydes and Ketones
J . Org. Chem., Vol. 61, No. 11, 1996 3851
secondary amines was successful under the standard
conditions and gave the desired products in good to
excellent yields. The scope of the reaction includes
different alicyclic ketones, from cyclobutanone to cy-
clododecanone (Table 1: entries 1-23), bicyclic ketones
such as norcamphor and tropinone (Table 1: entries 24-
30), saturated acyclic ketones (Table 1: entries 31-39),
and keto esters (Table 1: entries 47-49). Nearly all
primary and nonhindered secondary amines were used
successfully in these reactions. For the same ketone, the
rate of the reaction was dependent on the steric and
electronic factors associated with the amines. In general,
primary aliphatic amines reacted faster than primary
aromatic and secondary aliphatic amines (Table 1: en-
tries 10 vs 11; 14 and 15 vs 16 and 17; 24 vs 26 and 27).
Cyclic secondary amines such as morpholine reacted
faster than acyclic secondary amines such as diethyl-
amine (Table 1: entry 33 vs 36) while the sterically
hindered diisopropylamine did not react even after days
(Table 1: entry 45). In some slow reactions (e.g., Table
1: entries 11, 27, 32, 34, and 36), small amounts of side
products were formed (1-5% by GC area % analysis)
from N-ethylation and N-acetylation of the starting
amines.22 These impurities were easily removed in the
workup or during the recrystallization of the salts.
The reaction conditions are very mild and can tolerate
the presence of acid sensitive functional groups such as
acetals and ketals. For example, the reductive amination
of cyclohexanedione monoethylene ketal with primary
and secondary amines afforded good to excellent isolated
yields of the corresponding amines (Table 1: entries 14-
18). Another example is the reductive alkylation of
aminoacetaldehyde diethyl acetal with cyclododecanone
(Table 1: entry 9). A particularly useful example is the
reaction involving cyclohexanedione monoethylene ketal
and aminoacetaldehyde diethyl acetal (Table 1: entry 18)
which provides a secondary amine product containing
protected aldehyde and ketone functionalities in a nearly
quantitative yield.
Of all the ketones used in this study, small aliphatic
cyclic ketones, ranging from cyclobutanone to cyclohex-
anone, were most reactive. Larger cyclic ketones such
as cyclooctanone and cyclododecanone and acyclic ketones
such as 2-heptanone reacted somewhat slower. Reactiv-
ity of cyclobutanone was so high that its reaction with
benzylamine gave a mixture of mono- and dicyclobutyl-
benzylamines even with the use of excess amine (Table
1: entry 2). Clean formation of N,N-dicyclobutylbenzyl-
amine resulted with the use of a 1:2 molar ratio of amine
to ketone (Table 1: entry 1). The reactions with second-
ary amines were very effective since there was no
dialkylation product (Table 1: entry 3).
The least reactive ketones were aromatic, R,â-unsatur-
ated, and sterically hindered ketones. Aromatic and R,â-
unsaturated ketones reacted very slowly (Table 1: entries
40, 41, and 43). Experimentally, a saturated aliphatic
ketone was reductively aminated, selectively, and quan-
titatively in the presence of an aromatic or R,â-unsatur-
ated ketone (Table 1: entries 42 and 44). The unreacted
ketones were recovered unchanged except for the forma-
tion of trace amounts of their imines (as determined by
GC/MS analysis of the reaction mixture). Sterically
hindered ketones were even less reactive than aromatic
and R,â-unsaturated ketones, e.g., camphor showed no
reaction with benzylamine after four days (Table 1: entry
46). The steric factors associated with both ketones and
amines seem to be very important in determining the
outcome of the reaction.
In reactions where the formation of diastereomers was
possible, we observed varying degrees of selectivity.
Reductive amination of 4-tert-butylcyclohexanone with
pyrrolidine and cyclohexylamine occurred with a moder-
ate diastereoselectivity to give the thermodynamically
less favored cis products. This results from equatorial
attack by the hydride reagent on the intermediate imine,
to form the axial amine (Table 1: entries 19 and 20).23
The reductive amination of androstanolone with isopro-
pylamine gave a mixture of 3R- and 3â-(isopropylamino)-
androstan-17â-ol in about 25:75 ratio (Table 1: entry 21).
The reactions involving bicyclic ketones showed higher
degrees of diastereoselectivity. For example, the reduc-
tive aminations of norcamphor led to the exclusive
formation of the endo products, from exo attack by the
hydride reagent. The reductive amination of norcamphor
with benzylamine (Table 1: entry 24) produced a single
product. This product was identical to that obtained from
the reductive amination of benzaldehyde with endo-2-
aminonorbornane (Table 2: entry 13), thus confirming
the endo stereochemistry of the product.
Reductive amination of tropinone with primary amines
such as benzylamine and aniline (Table 1: entries 28 and
29) was accomplished in good yield and high diastereo-
selectivity giving the endo isomer as the major product
1
(determined by H NMR).24 The reaction with benzyl-
amine gave the endo and exo products in approximately
20:1 ratio while the reaction with aniline showed no
detectable exo product. The reaction of tropinone with
piperidine was extremely sluggish giving low conversion
to about a 1:1 mixture of the exo- and endo-products after
four days of reaction (Table 1: entry 30).
The poor solubility of ammonium acetate in DCE, THF,
or CH3CN limits its use in the reductive amination of
ketones to prepare primary amines. The initial primary
amine product is much more soluble than ammonium
acetate and reacts faster with the ketone to generate
dialkylamines, so this reaction can be used for the
preparation of symmetrical dialkylamines. The amina-
tion reaction is relatively slow and some ketone reduction
may occur if AcOH is added. Even the use of a large
excess (10 equiv) of ammonium acetate in THF, DCE, or
CH3CN, in the presence of Et3N, did not favor the
formation of the monoalkylamine, the only product
formed was dicycloheptylamine (Table 1: entries 22
and 23).
N-Substituted R-amino esters were prepared by the
reductive amination of R-keto esters with primary and
(23) This result is consistent with literature reports on the reduction
of 4-substituted cyclohexanone imines or iminium salts which con-
cluded that bulky hydride reagents such as L-Selectride attack
preferentially from the less hindered equatorial side to give the cis-
products, while less bulky hydride reagents such as NaBH4 and NaBH3-
CN slightly favor the axial approach, see: (a) Wrobel, J . E.; Ganem,
B. Tetrahedron Lett. 1981, 22, 3447. (b) Hutchins, R. O.; Markowitz,
M. J . Org. Chem. 1981, 46, 3571. (c) Hutchins, R. O.; Su, W.-Y.;
Sivakumar, R.; Cistone, F.; Stercho, Y. P. J . Org. Chem. 1983, 48, 3412.
(d) Hutchins, R. O.; Adams, J .; Rutledge, M. C. J . Org. Chem. 1995,
60, 7396.
(24) Chemical shift assignments of individual protons were arrived
at by COSY, HETCOR, and Inverse HMBC NMR experiments. The
stereochemistry of N-phenyl-3-aminotropane and N-benzyl-3-aminotro-
pane was assigned based on 1D NOE and coupling experiments. The
assignments were in line with other literature reports; cf. Bagley, J .
R.; Riley, T. N. J . Hetrocycl. Chem. 1977, 14, 599 and references
therein.
(22) The N-ethylation of amines is a major process in reaction of
amines with sodium borohydride in neat acetic acid and is believed to
proceed through an acetaldehyde formation.21a

3852 J . Org. Chem., Vol. 61, No. 11, 1996
Ta ble 1. Red u ctive Am in a tion of Keton es
Abdel-Magid et al.

Reductive Amination of Aldehydes and Ketones
J . Org. Chem., Vol. 61, No. 11, 1996 3853
Ta ble 1. (Con tin u ed )
secondary amines. The reductive amination of methyl
pyruvate with benzylamine was a fast and efficient
reaction under the standard conditions that gave N-
benzylalanine methyl ester in an excellent yield (Table
1: entry 47). The reaction was slower with hindered keto
esters such as methyl 3-methyl-2-oxobutanoate (Table
1: entry 48), and the competing ketone reduction was a
major reaction. The aromatic keto ester, methyl benzoyl
formate reacted even slower with benzylamine and was
also accompanied by considerable ketone reduction (Table
1: entry 49). Reactions involving other less reactive
amines such as aniline or morpholine were much slower

3854 J . Org. Chem., Vol. 61, No. 11, 1996
Abdel-Magid et al.
and gave increasing amounts of ketone reductions.
Whenever ketone reduction was a problem, the conditions
were modified to use the amines as limiting reagents.
The excess ketones were completely reduced to the
corresponding alcohols which required the occasional
addition of excess reducing agent. N-Substituted R-ami-
no esters were also prepared by the reductive amination
of R-amino esters with ketones, e.g., N-(2-butyl)glycine
ethyl ester was prepared in very good yield from ethyl
glycinate and 2-butanone (Table 1: entry 50).
The reductive amination of cyclohexanedione mono-
ethylene ketal with phenylhydrazine gave cleanly the
N-substituted phenylhydrazine (Table 1: entry 51) in
nearly quantitative yield. Other ketones such as cyclo-
hexanone and 2-heptanone reacted to give similar prod-
ucts (Table 1: entries 52 and 53); however, there were
some competing side reactions. The crude products
showed the formation of byproducts, about 15% with
cyclohexanone and 31% with 2-heptanone. Attempted
reductive amination of cyclooctanone with hydroxylamine
was not successful and resulted in the oxime formation
(Table 1: entry 54).
Reductive aminations in which diamines containing
both primary and secondary amino groups were studied
and in general, primary amines reacted faster. In the
case where the primary amino group was aliphatic and
the secondary group was aromatic, e.g., N-phenylethyl-
enediamine, there was a clear difference in reactivity
between the two amines. The reaction with 4-heptanone
gave a quantitative yield of the product resulting from
exclusive reaction with the primary amine (Table 1:
entry 55). In the case where both amino groups were
aliphatic, such as 1-(2-aminoethyl)piperazine, there was
a high selectivity (94:6) for the primary group in reaction
with acetylcyclohexane (Table 1: entry 56).
(b) Red u ctive Am in a tion of Ald eh yd es. Unlike
ketones, aldehydes can be reduced with sodium triac-
etoxyborohydride.18 Thus, the possibility exists that the
reduction of the aldehyde would compete with the reduc-
tive amination process under the standard conditions.
However, these conditions were so selective that the
reductive aminations with aldehydes occurred very ef-
fectively and resulted in fast reactions with no aldehyde
reductions in most cases studied (Table 2). One case in
which aldehyde reduction was detected involved a reac-
tion with the very sterically hindered diisopropylamine
(Table 2: entry 6). All other examples in Table 2 resulted
in fast and efficient reductive aminations with a variety
of aliphatic primary and secondary amines as well as
aniline with no detectable aldehyde reductions. Both
aliphatic and aromatic aldehydes were very reactive and
gave reductive amination products with a broad variety
of primary and secondary amines. The reaction times
ranged from 20 min to 24 h. The mildness of the reac-
tion conditions is well illustrated in the reductive ami-
nation of 1,1′,2-tris-nor-squalene aldehyde with dieth-
ylamine and diisopropylamine (Table 2: entries 19 and
20). The aldehyde was cleanly converted to the corre-
sponding amines in high yields with no detectable side
reactions.
certain substrates.25 An alternative stepwise procedure
for such systems is discussed later in this paper.
Some aldehydes, such as formaldehyde and glutaral-
dehyde are only available commercially as aqueous
solutions which may restrict their use under the above
reaction conditions because of the decomposition of the
hydride reagent with water. However, the reaction may
be carried out in DCE with excess hydride reagent, e.g.,
the reductive amination of either aqueous glutaraldehyde
or formaldehyde with 1-phenylpiperazine and about 4
hydride equivalents was carried out on 10 mmol scale
(Table 2: entries 21 and 22) and gave nearly quantitative
yields of the corresponding amines. This, however, may
not be suitable for large scale reactions.
The use of phenylhydrazine in reductive amination of
benzaldehyde was not successful, resulting only in hy-
drazone formation.
Generally, with either ketones or aldehydes, reactions
in DCE were noticeably faster than those carried out in
THF (e.g., Table 1: entries 6 vs 7; 25 vs 26 and Table 2:
entries 3 vs 4; 9 vs 10). Also, in the same solvent,
reactions were consistently faster in the presence of 1
(or more) mol equiv of acetic acid. For most ketones,
reactions were improved in the presence of acetic acid.
However, the addition of acetic acid is not always
advantageous to the reaction. Most reactions with al-
dehydes are fast and do not require addition of AcOH.
Addition of AcOH to a slow reaction, e.g., cyclohexan-
ecarboxaldehyde with diisopropylamine, resulted in a fast
reaction accompanied with about 25% aldehyde reduc-
tion, and the yield of the isolated desired product was
only 41%. When the reaction was carried out in the
absence of AcOH, the reaction was slower; however, the
aldehyde reduction was only about 5%, and the isolated
yield of the purified reductive amination product in-
creased to 75% (Table 2: entries 6 and 7). Direct com-
parisons were made between reactions in DCE and THF
and with or without added acetic acid in representative
reactions. The rate of product formation was determined
quantitatively26 in each case. The results of these com-
parisons were in agreement with the above observations.
(c) Red u ctive Am in a tion w ith Wea k ly Ba sic a n d
Non ba sic Am in es. Few literature references have dealt
with aromatic amines containing electron withdrawing
substituents in reductive amination reactions.13,21a Cata-
lytic hydrogenation conditions do not allow the presence
of many of the easily reduced electron-withdrawing
substituents such as cyano and nitro groups, since these
substituents are often reduced under catalytic hydroge-
nation conditions.6,7 On the other hand, we and others13a,b
have found the most used hydride reagent, sodium
cyanoborohydride [NaBH3CN], to be sluggish and inef-
ficient when used with these weak bases in reductive
amination reactions.
As a consequence of substitution by electron-withdraw-
ing substituents, these amines are both poor nucleophiles
and weak bases (e.g., pKa 3.98 for 4-chloroaniline, 1.02
for 4-nitroaniline, -0.29 for 2-nitroaniline, -4.26 for 2,4-
dinitroaniline).27 This slows the initial nucleophilic at-
In the reductive amination of aldehydes with primary
amines, formation of dialkylated amines is a common side
reaction.5 This side reaction was rarely a problem in
most reactions reported in Table 2. In the few cases
when it was detected, it was suppressed by the addition
of up to 5% molar excess of the primary amine. However,
the dialkylation of amines remained a problem with
(25) For a discussion of the dialkylation side reactions involving γ-
and δ-amino esters with aldehydes and a mechanistic explanation,
see: Abdel-Magid, A. F.; Harris, B. D.; Maryanoff, C. A. Synlett 1994,
81.
(26) The progress of these reactions was followed by GC. Linear
standard curves of the response factors by GC areas of starting
materials and expected products were determined to allow the
quantitative measurements of their concentrations in the reaction
mixtures.

Reductive Amination of Aldehydes and Ketones
J . Org. Chem., Vol. 61, No. 11, 1996 3855
Ta ble 2. Red u ctive Am in a tion of Ald eh yd es
tack on the carbonyl carbon and leads to slower overall
reaction rates (Scheme 2). In addition, the carbonyl
group now competes effectively with the less basic
intermediate imine for protonation and subsequently for
the hydride in the reduction step.2b This may lead to a
significant carbonyl reduction, consumption of both the
carbonyl compound and the reducing agent and low yields
of the reductive amination products. The reducing agent
and reaction conditions should be chosen carefully to
minimize such side reactions.

3856 J . Org. Chem., Vol. 61, No. 11, 1996
Abdel-Magid et al.
Sch em e 2
nostilbene reacted with hexanal to give a high yield of
the N-hexyl product, the dihydro analogue iminodiben-
zyl gave no reaction under the same conditions (Table 3:
entry 25).
One of the most unique reactions, however, was the
reductive alkylation of p-toluenesulfonamide with ben-
zaldehyde to give the N-benzyl derivative (Table 3: entry
26). The reaction is carried out initially under the
standard conditions in the presence of Et3N (2 equiv).
The aldehyde is usually consumed in about 24 h to give
a mixture of N-benzyl p-toluenesulfonamide and N-
benzal p-toluenesulfonamide. The reaction mixture is
then treated with AcOH (2.5 equiv) and additional
NaBH(OAc)3 (1 equiv) to finish the reduction. The
reaction, however, was not successful with ketones or
carboxamides.
The least reactive amines, 2,4-dinitroaniline and 2,4,6-
trichloroaniline failed to undergo reductive amination
with benzaldehyde (Table 3: entries 20 and 21). Cyclo-
hexanecarboxaldehyde, on the other hand, reacted slowly
with these two amines to give the corresponding reduc-
tive amination products. In these two reactions, the
aldehyde reduction became a major reaction process. To
assure the presence of enough aldehyde to react with the
amine, the reaction required occasional additions of
aldehyde and reducing agent, up to 5 equiv each and over
a two to four day period (Scheme 3). The reactions
progressed to reach 90-92% conversion (as determined
by GC) and gave 61% and 58% isolated yields, respec-
tively, after chromatography. It is possible that these
reactions proceed via initial formation of intermediate
enamines rather than imines which may explain the lack
of reactivity of aromatic aldehydes which cannot form
enamines.
Sodium triacetoxyborohydride is very efficient in re-
ductive amination reactions with such unreactive amines.
The results from several reactions are listed in Table 3.
In several cases such as monosubstituted anilines (e.g.,
p-nitro-p-carbethoxy-, and p-cyanoanilines), the standard
reaction conditions described previously (about 1:1 ratio
of carbonyl compound to amine, 1.4 equiv of NaBH(OAc)3
with 1 equiv of acetic acid) were adequate. However,
with less basic amines such as o-nitroaniline, 2,4-dichlo-
roaniline, or 2-aminothiazole, the reaction conditions
were modified to compensate for the aforementioned
effects and to maximize the yields of the reductive
amination products. The optimum conditions included
the use of the amine as the limiting reagent with 1.5-2
mol equiv of the carbonyl compound, 2-3 equiv of NaBH-
(OAc)3, and 2-5 equiv of AcOH in 1,2-dichloroethane.
Under these conditions, a variety of weakly basic amines
were successfully employed in the reductive aminations
of ketones and aldehydes in isolated yields ranging from
60% to 96%. The reaction is convenient and the condi-
tions are mild and show a high degree of tolerance for a
variety of functional groups including nitro, cyano, halo,
carboxy, and carbethoxy groups.
(d ) Com p a r ison w ith Oth er Red u cin g Agen ts. In
general, the results of reductive amination employing
sodium triacetoxyborohydride were as good as or better
than most comparable reported results whether done
using hydrogenation or hydride reagents. However, in
many cases, our results were far superior to others. For
example, we compared the reductive amination of cyclo-
hexanone with morpholine using NaBH3CN vs NaBH-
(OAc)3. The reaction with NaBH3CN (6 hydride equiv)
in methanol and in the absence of AcOH was only 34%
complete after 23 h with the formation of about 10% of
the corresponding enamine. The conversion improved to
50% in 23 h with AcOH (1 equiv) with no enamine
formation. The reaction using the standard NaBH(OAc)3
conditions was 99.8% complete in 3 h without a trace of
enamine formation (Table 1: entry 5). In another
comparison, the reductive amination of 1-carbethoxy-4-
piperidinone with p-chloroaniline was only 45% complete
with NaBH3CN after 22 h but was >96% with NaBH-
(OAc)3 in 2.5 h (Table 3: entry 4).
Ketones reacted effectively with p-monosubstituted
anilines to give good yields of the reductive amination
products (Table 3: entries 1-8). The reaction was
slightly slower with 2,4-dichloroaniline and gave a high
yield of the desired reductive amination product in
addition to some ketone reduction and the formation of
about 3% of N-ethyl-2,4-dichloroaniline (Table 3: entry
9). The reaction became very slow with o-nitroaniline
which progressed only to about 30% conversion to the
reductive amination product and 17% of N-ethyl-2-
nitroaniline after 6 days (Table 3: entry 10). The
reaction stopped completely when both ortho positions
were substituted as in 2,6-dibromo- and 2,4,6-trichloro-
anilines (Table 3: entries 11 and 12).
The reactions with aldehydes were faster than those
with ketones and gave higher yields from similar reac-
tions. Aldehyde reductions occurred only with the least
reactive amines. In the reductive amination of aldehydes
with p-carboxyaniline and p-nitroaniline (Table 3: en-
tries 13, 14, and 18), no competing aldehyde reduction
was observed. In these cases, the standard conditions
were used. With weaker amines such as 2,4-dichloro-
aniline and o-nitroaniline (Table 3: entries 15 and 16),
the conditions were modified to use the amine as a
limiting reagent since aldehyde reduction occurred to
the extent of 10-30%. This procedure was applied to
other weakly basic primary amines such as 2-aminothia-
zole (Table 3: entries 22 and 23) and secondary amines
such as iminostilbene (Table 3: entry 24). While imi-
An impressive result was obtained in the reductive
amination of 1,1′,2-tris-nor-squalene aldehyde with di-
ethylamine and isopropylamine. These reactions were
reported to give about 5% yield under regular Borch
conditions.28 The yields were improved to 46 and 42%,
respectively, when the reactions were carried out with
NaBH3CN in anhydrous THF in the presence of HCl (pH
3).29 Under our standard conditions, these reactions gave
(28) Duriatti, A.; Bouvier-Nave, P.; Benveniste, P.; Schuber, F.;
Delprino, L.; Balliano, G.; Cattel, L. Biochem. Pharmacol. 1985, 34,
2765.
(29) Ceruti, M.; Balliano, G.; Viola, F.; Cattel; L.; Gerst, N.; Schuber,
F. Eur. J . Med. Chem. 1987, 22, 199.
(27) (a) Albert, A.; Serjeant, E. P. In The Determination of Ionization
Constants; Chapman and Hall: London, 1971; p 91. (b) Yates, K; Wai,
H. J . Am. Chem. Soc. 1964, 86, 5408.

Reductive Amination of Aldehydes and Ketones
J . Org. Chem., Vol. 61, No. 11, 1996 3857
Ta ble 3. Red u ctive Am in a tion w ith Wea k ly Ba sic Am in es
98 and 90% yield of products in high purity without
chromatography (Table 2: entries 19 and 20).
amounts of reagents under the standard conditions
(Table 1: entries 12, 13, and 38).
Borane-pyridine is another reducing agent used for
reductive amination reactions.13a Its use in the reductive
amination of either pyridine-4-carboxaldehyde or m-
nitrobenzaldehyde with ethyl piperidine-2-carboxylate
The reductive amination of either 2-indanone, â-te-
tralone, or phenylacetone with aniline was reported to
work when aniline was used as solvent under catalytic
hydrogenation conditions and gave 72%, 54%, and 21%
yield of products, respectively.30 We obtained 85%, 87%,
and 80% yield of these products using stoichiometric
(30) Campbell, J . B.; Lavagnino, E. P. In Catalysis in Organic
Syntheses; J ones, W. H., Ed.; Academic Press: New York, 1980; p 43.

3858 J . Org. Chem., Vol. 61, No. 11, 1996
Abdel-Magid et al.
Sch em e 3
a very convenient and efficient alternative for carrying
out reductive aminations on substrates which tend to give
significant amounts of dialkylated products using the
direct procedure.
A similar transformation along the lines of prior
formation of the imine is prior formation and then
reduction of the enamine. We have found that sodium
triacetoxyborohydride reduces enamines quickly and
efficiently in DCE to give products in high yields (Table
5: entries 6 and 7), making it a useful reagent for this
stepwise procedure.
Another very efficient stepwise reductive amination
procedure was developed by Mattson et al.13b In this
procedure, the amine and the carbonyl compound are
mixed in neat titanium(IV) isopropoxide and the result-
ing intermediate is reduced with NaBH3CN in ethanol.
The authors reported the formation of a carbinol amine
intermediate (see Table 5: entries 8-10) rather than the
usual iminium ion. The method is applicable to primary
and secondary amines. We modified this procedure to
reduce the intermediate with NaBH4 in methanol (in-
stead of NaBH3CN/ethanol).31 This modified reduction
is very fast and gives the desired amine in very good yield
and high purity, e.g., the reductive amination of the
unsaturated ketone 1-acetylcyclohexene with benzyl-
lamine, which was very slow under our standard condi-
tions, proceeded very fast under these modified conditions
to give a quantitative yield of the reductive amination
product (Table 5: entry 8). Of particular interest, the
reductive amination of tropinone with benzylamine using
this procedure which gave a near quantitative yield of a
3:2 mixture of the endo- and exo products (Table 5: entry
10). As reported earlier, this reaction gave approximately
20:1 ratio of the endo and exo products using NaBH(OAc)3
(Table 1: entry 28). Attempts to achieve a reaction
between tropinone and aniline using this modified system
resulted in incomplete reactions. In regards to secondary
amines, Mattson et al.13b reported a successful reductive
amination of tropinone with piperidine in 58% yield;
however, the stereochemistry of the product was not
disclosed. Our modified conditions gave the product in
a 90% crude yield as a mixture of the endo- and exo-
products in about 1:7 ratio. The pure oxalate salt of the
exo product was isolated in 66% yield (Table 5: entry 9).
The exo-stereoselectivity of this reaction is opposite to
that obtained from the primary amines. We are examin-
ing these systems to better understand their mechanistic
pathway.
gave 12% and 13% yield of isolated reductive amination
products, respectively, and considerable aldehyde
reduction.14b Under our standard conditions, we did not
observe any aldehyde reduction, and the isolated yields
were nearly quantitative (Table 2: entries 17 and 18).
(e) Step w ise (In d ir ect) Red u ctive Am in a tion of
Ald eh yd es. Occasionally, some reactions of aldehydes
and primary amines gave considerable amounts of di-
alkylation or other side products. As we mentioned
previously, the addition of a slight excess of the primary
amine may suppress this side reaction. However, the
dialkylation of certain primary amines such as γ- or
δ-amino esters and cyclobutylamine, remained a problem.
This also occurred when certain aldehydes were used,
such as cinnamaldehyde, hydrocinnamaldehyde, and
some straight chain aldehydes. Those particular cases
gave mixtures of monoalkylated and dialkylated amines
in ratios ranging from 5:1 and up to 1:1 under the
standard reaction conditions. Surprisingly, in some
reactions, the dialkylation side reaction happened even
when excess amine was used or when the reaction was
carried out with a “preformed imine” and no excess
aldehyde present. A discussion and a possible explana-
tion of some of these results in case of γ- and δ-amino
esters was reported.25 We developed an alternative
procedure for such reactions which involves the fast
reduction of preformed imines. Imine formation, how-
ever, is a reversible reaction and requires long reaction
times and the use of a dehydrating agent such as
molecular sieves or azeotropic removal of water to drive
the reaction to completion. We studied the relative rates
of imine formation from aldehydes and primary amines
in different solvents, namely THF, DCE, and methanol
without added catalysts or dehydrating agents. A com-
parison of the relative rates of imine formation in the
three solvents is listed in Table 4. The relative ratio of
product imine to reactant aldehyde was determined by
GC analysis of the reaction mixture and confirmed in
Su m m a r y a n d Con clu sion s
The results presented here indicate that sodium tri-
acetoxyborohydride is a synthetically useful reagent for
reductive aminations of aldehydes and ketones. It is a
mild, commercially available reagent, and it is a reagent
of choice for reductive amination of carbonyl compounds.
The scope of the reaction covers all aldehydes and
unhindered aliphatic ketones. Aliphatic ketones (and
aldehydes) can be selectively reductively aminated in the
1
some cases by H NMR in CD3OD, THF-d8, or CDCl3. In
all the cases listed in Table 4, reactions in methanol were
consistently faster and gave nearly quantitative conver-
sions relative to those carried out in THF or DCE. The
imine formation was slower for ketones in all three
solvents; however, the reaction was still faster in metha-
nol (Table 4: entry 7).
The reduction of the aldimines was carried out directly
in methanol with sodium borohydride to give the corre-
sponding secondary amines in very high yields and very
short reaction times. Thus, this stepwise (or indirect)
one-pot procedure involving imine formation in methanol
followed by in situ reduction with sodium borohydride is
(31) (a) We reported the use of this modification of Mattson’s
procedure (Ti(OiPr)4-NaBH4/MeOH, rt) at the 198th ACS National
Meeting, Miami Beach, FL, September 1989; paper ORGN 154 and at
The International Chemical Congress of Pacific Basin Societies,
Honolulu, HI, December 1989; paper ORGN 409, Abdel-Magid, A. F.;
Maryanoff, C. A.; Sorgi, K. L.; Carson, K. G. A similar modification
[Ti(OiPr)4-NaBH4 in diglyme at 60 °C or in ethanol at 25 °C)] was
reported recently: (b) Bhattacharyya, S. Tetrahedron Lett. 1994, 35,
2401. (c) Bhattacharyya, S. Synlett 1994, 1029; (d) Bhattacharyya, S.
Synth. Commun. 1995, 25, 9.

Reductive Amination of Aldehydes and Ketones
J . Org. Chem., Vol. 61, No. 11, 1996 3859
Ta ble 4. Im in e F or m a tion fr om Ald eh yd es a n d P r im a r y Am in es
Yield (%)
1
Entry
1
Aldehyde
Amine
Time (h)
by GC (solvent)
by H NMR (solvent)
benzaldehyde
tert-BuNH2
aniline
22
22
4
4.2
4.3
1.5
24
26
2.4
2
0.24
0.25
5.4
5.3
2.7
4.6
5.3
2.7
23
84 (DCE)
95 (THF)
97 (MeOH)
81 (DCE)
74 (THF)
99 (MeOH)
92 (DCE)
84 (THF)
99 (MeOH)
97 (DCE)
95 (THF)
98 (MeOH)
89 (DCE)
90 (THF)
96 (MeOH)
89 (DCE)
90 (THF)
94 (MeOH)
53 (DCE)
75 (THF)
85 (MeOH)
80 (CDCl3)
98 (THF-d8)
97 (CD3OD)
90 (CDCl3)
75 (THF-d8)
97 (CD3OD)
2
3
4
5
6
7
benzaldehyde
m-anisaldehyde
m-anisaldehyde
c-C6H11CHO
aniline
2-(3,4-dimethoxyphenyl)ethylamine
100 (CDCl3)
100 (THF-d8)
100 (CD3OD)
aniline
c-C6H11CHO
tert-BuNH2
benzylamine
cycloheptanone
23
23
Ta ble 5. Step w ise Red u ctive Am in a tion Rea ction s
presence of aromatic and R,â-unsaturated ketones. Ali-
phatic and aromatic primary and sterically unhindered
secondary amines can be used in these reactions. The
procedure was superior with weakly basic and nonbasic
amines. In representative comparisons with other com-
monly used reducing reductive amination reagents,
sodium triacetoxyborohydride reacted consistently faster,
gave better yields, and produced fewer side products.
Methanol allows a rapid formation of imines from alde-
hydes and primary amines. Aldimines formed in metha-
nol were efficiently reduced to their corresponding amines
using NaBH4.
Exp er im en ta l Section
Melting points are uncorrected. 1H NMR and 13C NMR
spectra were recorded at 400 MHz. The chemical shifts are
expressed as δ units with Me4Si as the internal standard. IR
spectra were recorded on an FT-IR spectrometer and absorp-
tions are reported in wave numbers (cm-1). GC-MS (EI) were
recorded on a GC/MSD instrument using a cross linked methyl

3860 J . Org. Chem., Vol. 61, No. 11, 1996
Abdel-Magid et al.
silicone 12 m × 0.2 mm × 0.33 mm capillary column. Mass
spectra were recorded as CI or FAB. Accurate mass measure-
ments were carried out using a double-focusing instrument of
EB configuration (where E is an electric sector and B is a
magnet). The reported accurate mass (mass-to-charge ratio)
measurement values (mean ( 3σ) are for the [MH]+ ions of
the analytes of interest and are the average of nine individual
determinations. The deviation of the experimental determina-
tions from the theoretical values are expressed in parts-per-
million (ppm). GC analyses were carried out on a cross linked
methyl silicone 12 m × 0.2 mm × 0.33 µm capillary column or
DB-17; 15 m × 0.2 mm × 0.25 µm. Sodium triacetoxyboro-
hydride was purchased from Aldrich Chemical Company, Inc.
Most reagents were commercially available reagent grade
chemicals and used without further purification.
P r oced u r es. 1. Dir ect Red u ctive Am in a tion Meth od s.
Gen er a l Notes. (a) The amine and the carbonyl compound
are mixed in 1,2-dichloroethane and treated with NaBH(OAc)3.
THF, CH2Cl2, or CH3CN may also be used as solvents.
(b) Acetic acid (1-2 mol equiv) may be used in reactions of
ketones but is not necessary with most aldehydes.
(c) Reactions are normally carried out using the free amines;
however, the amine salt may be used. In this case, 1-2 equiv
of Et3N is added to the reaction mixture. The Et3N must be
removed from basified product prior to salt formation.
(d) The progress of the reaction is followed by GC analysis
in most cases. A small aliquot is withdrawn from the reaction
mixture, quenched with aqueous NaOH or aqueous NaHCO3
and extracted with ether or any other suitable solvent and
injected into the GC. In case of high MW or heat-sensitive
compounds HPLC or TLC could be used.
(e) The reactions are usually quenched with aqueous 1 N
NaOH. In the presence of esters, or other alkali hydroxide
sensitive functional groups, aqueous NaHCO3 is used for
quench. Reactive amines such as benzylamine may dissolve
in aqueous NaHCO3 and that may give a false indication of
their consumption.
(f) Most amines form HCl salts cleanly in ether with ethereal
HCl, and most of these salts are recrystallized from EtOAc/
MeOH. Some aromatic HCl salts may turn dark and a few
aliphatic amine salts may be difficult to crystallize or may be
hygroscopic; in these cases the oxalate salts are prepared,
usually in MeOH. Picrate salts are prepared in ethanol.
Meth od I. This procedure is used for most ketone reactions.
A representative example is the reductive amination of cyclo-
pentanone with hexamethyleneimine (Table 1: entry 4):
Hexamethyleneimine (1.0 g, 10 mmol) and cyclopentanone
(0.84 g, 10 mmol) were mixed in 1,2-dichloroethane (35 mL)
and then treated with sodium triacetoxyborohydride (3.0 g,
14 mmol) and AcOH (0.6 g, 10 mmol). The mixture was stirred
at rt under a N2 atmosphere for 24 h until the reactants were
consumed as determined by GC analysis. The reaction
mixture was quenched by adding 1 N NaOH, and the product
was extracted with ether. The ether extract was washed with
brine and dried (MgSO4). The solvent was evaporated to give
the crude free base (1.60 g, 96%). The oxalate salt was
prepared in EtOAc/MeOH as shiny white crystals (2.2 g,
85.5%): mp 171-172 °C; FT-IR (KBr) 3435 (w), 2934 (s), 2872
(m), 2683 (m), 2648 (m), 2585 (m), 2524 (m), 1720 (m), 1623
(m), 1455 (m), 1404 (m), 1204 (m), 1116 (m), 717 (m), 499 (m),
466 (m) cm-1; 1H NMR (free base, CDCl3) δ 2.94-2.83 (m, 1H),
2.68-2.65 (m, 4H), 1.87-1.25 (m, 16H); 13C NMR (free base,
CDCl3) δ 65.9 (CH), 53.6 (CH2), 30.2 (CH2), 27.9 (CH2), 26.9
(CH2), 24.1 (CH2); EI MS m/ z (relative intensity) 167 (M+, 15),
138 (100), 124 (17), 110 (24), 96 (9), 82 (7), 55 (16). Anal. Calcd
for C13H23NO4: C, 60.68; H, 9.01; N, 5.44. Found: C, 60.69;
H, 9.03; N, 5.44.
sodium triacetoxyborohydride (3.0 g, 14 mmol). The mixture
was stirred at rt under a N2 atmosphere for 1.5 h. The GC
analysis showed >90% conversion in 30 min, but the remain-
der of the time was needed for the complete conversion. The
reaction mixture was quenched by adding aqueous saturated
NaHCO3, and the product was extracted with EtOAc. The
EtOAc extract was dried (MgSO4), and the solvent was
evaporated to give the crude free base (2.40 g, 96.7%), >98%
pure by area % GC analysis. A small portion was converted
to the dioxalate salt: an off-white solid, mp 109-111 °C
(EtOAc/MeOH); FT-IR (KBr) 3076 (w), 2968 (w), 2870 (w), 2521
(m), 1738 (s), 1603 (s), 1504 (m), 1460 (w), 1372 (w), 1204 (s),
999 (w), 714 (m) cm-1; 1H NMR (free base, CDCl3) δ 8.52 (d, J
) 5.2 Hz, 2H), 7.29 (d, J ) 5.2 Hz, 2H), 4.19 (q, J ) 7.1 Hz,
2H), 3.80 (d, J ) 14.4 Hz, 1H), 3.43 (d, J ) 14.4 Hz, 1H), 3.20
(m, 1H), 2.90 (m, 1H), 2.20 (m, 1H), 1.85 (m, 2H), 1.55 (m,
3H), 1.45 (m, 1H), 1.28 (t, J ) 7.1 Hz, 3H); 13C NMR (free base,
CDCl3) δ 174.1 (C), 153.3 (C), 149.7 (CH), 125.8 (CH), 64.8
(CH), 61.8 (CH2), 59.9 (CH2), 51.0 (CH2), 30.5 (CH2), 26.4 (CH2),
23.1 (CH2), 15.8 (CH3); EI MS m/ z (relative intensity) 175 (M+
- CO2Et, 100), 147 (4), 93 (5), 92 (25), 82 (5), 65 (15). Anal.
Calcd for C18H24N2O10: C, 50.47; H, 5.65; N, 6.54. Found: C,
50.26; H, 5.74; N, 6.44.
Meth od III. Same as Method I except for using THF as a
solvent. An example is the reductive amination of 2-heptanone
with morpholine (Table 1: entry 33): 2-Heptanone (2.28 g,
20 mmol) and morpholine (1.92 g, 22 mmol) were mixed in
THF (80 mL) at rt under N2. Sodium triacetoxyborohydride
(6.36 g, 30 mmol) and glacial AcOH (1.20 g, 20 mmol) were
added, and the mixture was stirred at rt for 27 h. The reaction
mixture was quenched with aqueous saturated NaHCO3
solution, and the product was extracted with Et2O. The Et2O
extract was dried (MgSO4) and cooled in an ice bath and then
treated with ethereal HCl. The product precipitated as a white
solid. The solid was collected by filtration and dried; yield:
3.24 g, 73% (>98% by GC area % analysis) mp 150-151 °C.
The analytical sample was obtained as a white crystalline solid
by recrystallization from EtOAc/MeOH: mp 151-153 °C; FT-
IR (KBr) 3410 (m), 2924 (s), 2863 (s), 2656 (s), 2604 (s), 2462
(m), 1434 (m), 1266 (m), 1114 (s), 1073 (m), 928 (m), 881 (m);
1H NMR (CDCl3) δ 12.60 (bs, 1H), 4.43 (dt, J ) 12.8, 2.1 Hz,
2H), 3.98 (dd, J ) 12.8, 3.0 Hz, 2H), 3.30-3.16 (m, 3H), 3.09-
2.95 (m, 2H), 2.15-2.06 (m, 1H), 1.64-1.10 (m, 10H), 1.44 (d,
J ) 6.7 Hz, 3H), 0.90 (t, J ) 6.5 Hz, 3H); 13C NMR (CDCl3) δ
63.43 (CH2), 63.37 (CH2), 62.5 (CH), 47.9 (CH2), 47.4 (CH2),
31.2 (CH2), 30.0 (CH2), 25.8 (CH2), 22.2 (CH2), 13.7 (CH3), 13.4
(CH3); EI MS m/ z (relative intensity) 185 (M+, 5), 170 (28),
115 (31), 114 (100), 86 (7), 84 (7), 70 (28). Anal. Calcd for
C11H24ClNO: C, 59.58; H, 10.91; N, 6.32; Cl, 15.99. Found:
C, 59.66; H, 11.00; N, 6.42; Cl, 16.09.
Meth od IV. Same as method II except for using THF as a
solvent. An example is the reductive amination of â-tetralone
with cyclohexylamine (Table 1: entry 10): Cyclohexylamine
(1.09 g, 11 mmol) and â-tetralone (1.46 g, 10 mmol) were mixed
in THF (50 mL) at rt under N2. The mixture changed in about
1 min from yellow to dark blue in color. Sodium triacetoxy-
borohydride (3.18 g, 15 mmol) was added and the mixture
stirred at rt under a N2 atmosphere for 24 h. The color of the
reaction mixture became light pink at the end of the reaction
time. The reaction mixture was quenched by adding aqueous
3 N NaOH (the color changed to dark green), and the product
was extracted with Et2O. The Et2O extract was dried (MgSO4)
and cooled in an ice bath and then treated with ethereal HCl.
The product precipitated as a white solid with a light pink
shade. The solid was collected by filtration and dried; yield:
2.27 g, 85% (>98% by GC area % analysis). The analytical
sample was obtained as a white solid by recrystallization from
EtOAc/MeOH: mp 230-233 °C; FT-IR (KBr) 3415 (s), 2941
(s), 2819 (s), 2679 (m), 2506 (w), 2430 (w), 1452 (m), 742 (m)
Meth od II. The above procedure is followed without the
addition of glacial acetic acid. The reaction mixture remains
cloudy throughout the reaction. This procedure is more
appropriate with most aldehydes and unhindered aliphatic
ketones. A representative example is the reductive amination
of 4-pyridinecarboxaldehyde with ethyl 2-piperidinecarboxylate
(Table 2: entry 17): Ethyl-2-piperidinecarboxylate (1.57 g, 10
mmol) and 4-pyridinecarboxaldehyde (1.07 g, 10 mmol) were
mixed in 1,2-dichloroethane (35 mL) and then treated with
cm-1 1H NMR (CDCl3) δ 9.60 (bs, 2H), 7.16-7.02 (m, 4H),
;
3.62-3.45 (m, 1H), 3.43-3.12 (m, 3H), 2.91-2.80 (m, 2H),
2.61-2.50 (m, 1H), 2.40-2.17 (m, 3H), 1.91-1.60 (m, 6H),
1.39-1.18 (m, 3H); 13C NMR (CDCl3) δ 134.6 (C), 132.4 (C),
129.2 (CH), 128.6 (CH), 126.5 (CH), 126.1 (CH), 54.2 (CH), 51.0
(CH), 32.3 (CH2), 29.5 (CH2), 28.9 (CH2), 27.9 (CH2), 26.0 (CH2),
24.7 (CH2); EI MS m/ z (relative intensity) 229 (M+, 46), 187

Reductive Amination of Aldehydes and Ketones
J . Org. Chem., Vol. 61, No. 11, 1996 3861
(13), 186 (85), 158 (12), 132 (11), 131 (77), 130 (100), 129 (34),
128 (20), 116 (14), 115 (26), 104 (23), 100 (17), 91 (35), 78 (11),
77 (12), 56 (28), 55 (31). Anal. Calcd for C16H24ClN: C, 72.29;
H, 9.10; N, 5.27; Cl, 13.34. Found: C, 72.41; H, 9.27; N, 5.17;
Cl, 13.53.
clobutyl-4-chlorobenzylamine (Table 5: entry 4): p-Chloroben-
zaldehyde (1.40 g, 10 mmol) and cyclobutylamine (0.75 g, 10.6
mmol) were mixed in MeOH (40 mL) at rt under a N2
atmosphere. The mixture was stirred at rt for 3 h, until the
aldimine formation was completed (determined by GC). The
aldimine in MeOH was carefully treated with solid NaBH4 (0.6
g, 16 mmol). The reaction mixture was stirred for 10 min and
quenched with 1 M NaOH. The product was extracted with
ether. The ether extract was washed with saturated aqueous
NaCl and dried (MgSO4). The solvent was evaporated to give
the crude product as a nearly colorless oil (1.92 g, 98%) which
was >97% pure by area % GC analysis. The oil was dissolved
in ether and treated with ethereal HCl to give the HCl salt
which was recrystallized from EtOAc/MeOH as white crystals,
2.07 g, 89%, mp 237-238 °C; FT-IR (KBr) 2917 (s), 2804 (s),
2762 (s), 2435 (m), 1497 (m), 1432 (m), 1092 (m), 808 (m), 520
Meth od V. This method is similar to methods I and II
except for the use of acetonitrile as a solvent.
Meth od VI. This method was used with those reactions
involving hindered R-keto esters or weakly basic amines in
which a competitive reduction of the carbonyl compounds
occur. The amine is used as a limiting reagent. A representa-
tive example of the R-keto esters is the reductive amination
of methyl 3-methyl-2-oxobutanoate with benzylamine (Table
1: entry 48): methyl-3-methyl-2-oxobutanate (1.8 g, 12.48
mmol) and benzylamine (0.446 g, 4.16 mmol) in DCE (14 mL)
was treated with sodium triacetoxyborohydride (1.76 g, 8.32
mmol) at 0 °C. After stirring 21 h, GC analysis showed total
consumption of the benzylamine with the presence of the
product amine and the intermediate imine in an 8:1 relative
ratio. The reaction was then treated with additional sodium
triacetoxyborohydride (0.88 g, 4.16 mmol). After stirring an
additional 21 h, GC analysis showed completion of the reaction.
The reaction mixture was diluted with EtOAc (20 mL) and
quenched with distilled water (10 mL). After further dilution
with EtOAc (20 mL), the pH of the water layer was adjusted
to 7 with saturated aqueous NaHCO3. The EtOAc layer was
separated and dried (MgSO4). The solvent was removed under
reduced pressure to give the crude product (2.45 g). The crude
product was dissolved in ether and treated with ethereal HCl
to give the HCl salt, 0.92 g, 81% yield, as a white solid: mp
190-191 °C; FT-IR (KBr) 2966 (m), 2806 (m), 2696 (m), 1742
(s), 1565 (m), 1469 (m), 1424 (m), 1249 (s), 1030 (m), 753 (m),
704 (m) cm-1; 1H NMR (CDCl3) δ 10.65 (bs, 1H), 10.00 (bs 1H),
7.66-7.63 (m, 2H), 7.41-7.38 (m, 3H), 4.40 (d, J ) 13.3 Hz,
1H), 4.25 (d, J ) 13.3 Hz, 1H), 4.20 (q, J ) 7.1 Hz, 2H), 3.45-
3.35 (m, 1H), 2.75-2.60 (m, 1H), 1.30 (t, J ) 7.1 Hz, 3H), 1.15
(d, J ) 6.9 Hz, 3H), 1.09 (d, J ) 6.9 Hz, 3H); 13C NMR (CDCl3)
δ 166.7 (C), 130.9 (CH), 129.5 (CH), 129.0 (CH), 63.1 (CH),
62.2 (CH2), 50.3 (CH2), 29.4 (CH), 19.5 (CH3), 17.6 (CH3), 14.0
(CH3); CI MS m/ z (rel intensity) 236 (MH+, 100), 162 (38), 91
(20). Anal. Calcd for C14H22ClNO2: C, 61.87; H, 8.16: N, 5.15.
Found: C, 61.94; H, 8.21: N, 5.05.
(m) cm-1 1H NMR (free base, CDCl3) δ 7.25-7.22 (m, 4H),
;
3.64 (s, 2H), 3.28-3.21 (m, 1H), 2.23-2.16 (m, 2H), 1.71-1.63
(m, 4H), 1.38 (bs, 1H); 1H NMR (HCl salt, CDCl3 + CD3OD) δ
9.70 (bs, 1H), 7.50 (d, J ) 8.4 Hz, 2H), 7.37 (d, J ) 8.4 Hz,
2H), 3.94 (s, 2H), 3.52 (q, J ) 8.15 Hz, 1H), 2.95 (bs, 2H), 2.41-
2.35 (m, 2H), 2.22-2.13 (m, 2H), 1.99-1.89 (m, 1H), 1.83-
1.73 (m, 1H); 13C NMR (free base, CDCl3) δ 138.9 (C), 132.4
(C), 130.1 (CH), 129.4 (CH), 50.2 (CH), 53.4 (CH2), 30.9 (2 CH2),
14.6 (CH2); EI MS m/ z (relative intensity) 196 (M+), 167 (48),
125 (100), 89 (26), 39, (18). Anal. Calcd for C11H15Cl2N: C,
56.91; H, 6.51; N, 6.03; Cl, 30.54. Found: C, 56.89; H, 6.45;
N, 5.87; Cl, 30.82.
Red u ction of En a m in es. An example is the reduction of
1-morpholino-1-cyclohexene (Table 5, entry 6 and Table 1,
entry 5): 1-Morpholino-1-cyclohexene (1.67 g, 10 mmol) in
DCE (30 mL) and AcOH (0.61 g, 10 mmol) was treated with
sodium triacetoxyborohydride (3.0 g, 14 mmol) and stirred
under argon. The GC analysis determined that the reduction
was complete after 10 min. The reaction was quenched by
adding 1 M NaOH, and the product was extracted with ether.
The ether extract was dried (MgSO4), and the solvent was
evaporated under reduced pressure to give the crude product
as a colorless oil (1.70 g). The oil was dissolved in ether (50
mL), cooled in an ice bath, and treated with ethereal HCl to
give the HCl salt as a white solid. The solid was collected by
filtration air-dried, and then recrystallized from EtOAc/MeOH
to give the purified product (1.88 g, 91.4%): mp 260-262 °C.
Picrate salt: a yellow solid, mp 174-175 °C (ethanol) (lit.11
176-177 °C); FT-IR (HCl salt, KBr) 3425 (s), 2937 (s), 2861
(m), 2672 (s), 2606 (s), 2478 (m), 1448 (m), 1399 (w), 1267 (w),
1110 (s), 1070 (w), 950 (w) cm-1; 1H NMR (free base, CDCl3) δ
3.74 (t, J ) 5.0 Hz, 4H), 2.43 (t, J ) 5.0 Hz, 4H), 2.29 (m, 1H),
1.88 (d, J ) 10.0 Hz, 4H), 1.61 (d, J ) 13.0 Hz, 1H), 1.28 (m,
5H); 13CNMR (free base, CDCl3) δ 66.9 (CH2), 63.4 (CH), 49.3
(CH2), 28.4 (CH2), 25.9 (CH2), 25.4 (CH2); EI MS m/ z (relative
intensity) 169 (M+, 16), 127 (11), 126 (100), 98 (7), 83 (10), 82
(12), 68 (10), 56 (19), 55 (36). Anal. Calcd for C10H20ClNO:
C, 58.38; H, 9.80; N, 6.81; Cl, 17.23. Found: C, 58.22; H, 9.69;
N, 6.72; Cl, 17.25.
Th e Use of Tita n iu m (IV) Isop r op oxid e. An example is
the synthesis of N-[1-(1-cyclohexenyl)ethyl]benzylamine (Table
5, entry 8): 1-Acetylcyclohexene (1.24 g, 10 mmol) and
benzylamine (1.2 g, 11.2 mmol) were mixed in neat titanium-
(IV) isopropoxide (4.78 g, 16.8 mmol) and stirred under
nitrogen for 3 h. Methanol (45 mL) was added followed by
careful addition of NaBH4 (0.6 g, 16 mmol). Analysis of the
reaction mixture after 5 min by GC indicated a complete
reduction to the amine. The reaction was quenched by adding
0.1 N NaOH. The resulting mixture was filtered through
Celite, and the residue was washed with ether (2 × 50 mL)
and with CH2Cl2 (50 mL). The organic layer was separated
and dried (MgSO4). The solvent was removed under reduced
pressure to give the crude product (2.2 g). The crude product
was dissolved in ether and treated with ethereal HCl to give
the HCl salt as a white solid. The solid was purified by
recrystallization from EtOAc/MeOH to give 2.1 g, 83%, of white
crystals: mp 215-217 °C; FT-IR (KBr) 3417 (m), 2932 (vs),
2835 (m), 2782 (s), 2732 (s), 2481 (w), 2422 (w), 1581 (m), 1454
(m), 1381 (w), 748 (m), 680 (m) cm-1; 1H NMR (CDCl3) δ 9.65
(bs, 1H), 7.61 (d, J ) 7.4 Hz, 2H), 7.35 (t, J ) 7.4 Hz, 2H),
An example of the use of weakly basic amines is the
reductive amination of 1-Carbethoxy-4-piperidone with p-
nitroaniline (Table 3: entry 5): 1-Carbethoxy-4-piperidone (3.2
g, 19 mmol), p-nitroaniline (1.4 g, 10 mmol), and glacial AcOH
(3.6 g, 60 mmol) were mixed in 1,2-dichloroethane (45 mL).
Sodium triacetoxyborohydride (6.0 g, 28 mmol) was added to
the above solution and the reaction mixture stirred at room
temperature under N2 for 18 h (GC analysis indicated a
complete reaction in addition to ca 20% ketone reduction). The
reaction was quenched with saturated aqueous NaHCO3, and
the product was extracted with EtOAc (3 × 75 mL). The
EtOAc extract was dried (MgSO4), and the solvent was
evaporated to give a yellow semisolid (4.7 g). The semisolid
was triturated with ether/hexane (7:3) to disolve the excess
piperidone and the piperidol byproduct. The yellow solid was
collected by filtration and dried (1.75 g, 60%), >99% pure by
GC area % analysis. The analytical sample was obtained by
recrystallization from EtOAc/hexane: mp 172-174 °C; FT-IR
(KBr) 3327 (s), 3182 (w), 3097 (w), 2929 (w), 2870 (m), 2398
(m), 1679 (s), 1600 (s), 1546 (m), 1460 (s), 1387 (m), 1296 (s),
1
1234 (m), 1184 (m), 1144 (m), 1105 (s), 1033 (m), 937 (m); H
NMR (CDCl3/d6-DMSO) δ 8.03 (d, J ) 9.0 Hz, 2H), 6.59 (d, J
) 9.0 Hz, 2H), 6.14 (d, J ) 7.5 Hz, 1H), 4.16-4.10 (m, 4H),
3.58-3.50 (m, 1H), 3.02 (m, 2H), 2.04-2.00 (m, 2H), 1.51-
1.42 (m, 2H), 1.27 (t, J ) 7.0 Hz, 3H); 13C NMR (CDCl3/d6-
DMSO) δ 154.9 (C), 152.5 (C), 136.4 (C), 125.9 (CH), 110.6
(CH), 60.8 (CH2), 49.0 (CH), 41.9 (CH2), 31.0 (CH2), 14.2 (CH3);
MS (EI), 293 (95), 277 (15), 276 (76), 264 (9), 248 (12), 220
(23), 177 (19), 156 (29), 155 (100), 154 (9), 131 (10), 130 (23),
129 (16), 128 (13), 127 (37), 126 (41), 117 (14), 100 (21), 96
(19), 82 (28), 57 (18), 56 (57). Anal. Calcd for C14H19N3O4: C,
57.33; H, 6.53; N, 14.33. Found: C, 57.27; H, 6.54; N, 14.21.
2. Step w ise P r oced u r es. Ald im in e F or m a tion /Red u c-
tion in Meth a n ol. An example is the preparation of N-cy-

3862 J . Org. Chem., Vol. 61, No. 11, 1996
Abdel-Magid et al.
7.29 (d, J ) 7.0 Hz, 1H), 5.73 (s, 1H), 3.95-3.92 (m, 1H), 3.78-
3.72 (m, 1H), 3.44 (bs, 1H), 2.25-2.21 (m, 1H), 2.05-1.95 (m,
3H), 1.72-1.51 (m, 4H), 1.52 (d, J ) 6.8 Hz, 3H); 13C NMR
(CDCl3) δ 132.6 (C), 130.6 (CH), 130.5 (C), 129.9 (CH), 128.9
(CH), 128.7 (CH), 59.0 (CH), 48.3 (CH2), 25.1 (CH2), 23.8 (CH2),
22.2 (CH2), 21.9 (CH2), 17.4 (CH3); EI MS m/ z (relative
intensity) 215 (M+, 4), 201 (16), 200 (83), 134 (17), 109 (4),
108 (4), 105 (10), 93 (5), 92 (13), 91 (100), 79 (13), 77 (11). Anal.
Calcd for C15H22ClN: C, 71.55; H, 8.81; N, 5.56; Cl, 14.08.
Found: C, 71.73; H, 8.92; N, 5.40; Cl, 13.96.
the high resolution mass spectra. We also thank Dr.
Norman Santora, the chemical information specialist,
for his valuable help with the library search. We thank
professor Robert O. Hutchins (Drexel University) for
helpful discussions and for supplying some authentic
samples.
Su p p or tin g In for m a tion Ava ila ble: Supporting data for
nearly all products are available including mp, FT-IR, 1H
NMR, 13C NMR, MS, and elemental analysis (39 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
Ack n ow led gm en t. We are grateful to the PRI
spectroscopy group for their help and assistance: D.
Gauthier, G. Leo, and P. McDonnell for NMR spectra,
J . Masucci and W. J ones for the FT-IR and mass
spectra, and X. Xiang, D. Burinsky, and R. Dunphy for
J O960057X