This file is a part of the Rhodium site archive. This Aug 2004 static snapshot is hosted by Erowid
as of May 2005 and is not being updated. > > Back to Rhodium Archive Index > >

Reduction of Imines and Cleavage of Oximes by Sodium Dithionite

by Peter M. Pojer, Aust. J. Chem., 32, 201-4 (1979)
[ Back to the Chemistry Archive ]

Sodium dithionite reduces a variety of imines to secondary amines. Oximes are cleaved by the reagent to the parent carbonyl compound.


Sodium dithionite is a powerful, inexpensive, safe and readily available reducing agent. It has been used for more than 70 years in the reduction of aromatic nitro compounds [1], diazonium salts [2], a variety of pyridinium compounds [3,4], some complex oximes [5,6] and other nitrogen-containing functional groups [7]. Recently, two reports [4,8] appeared on the reduction of aldehydes and ketones with hot solutions of sodium dithionite.

It is surprising, therefore, that no investigations of its use in the reductive alkylations of amines have been described. In a related reaction, Hawthorne and coworkers [9] reported that sodium dithionite converts the diimines of biphenyl-2,2'-dicarbaldehyde into 6-substituted 6,7-dibydro-5H-dibenz[c,e]azepines. This group [9] also described the reduction of biphenyl-2,2'-dicarbaldehyde monooxime to 6,7-dihydro-5H-dibenz[c,e]azepine. Sodium dithionite has also been employed in the Knorr pyrrole synthesis where the intermediate ethyl 2-hydroxyiminoacetoacetate was reduced to an amine (which was not isolated) [5] and in the reduction of 2,3,5-trimethyl-1,4-benzoquinone I-oxime to 2,3,6-trimethyl-4-aminophenol [6]. Both of these oximes are highly conjugated and are therefore expected to be susceptible to reduction.

The use of metal hydrides and catalytic hydrogenation in the reduction of imines [10] and oximes, and derivatives of oximes [11] is well documented but many of the procedures suffer from the expensiveness or inconvenience of the reagents.

Reductive Alkylation of Amines

We found methanol/water [4,9] or dioxane/water [8] mixtures unsuitable for the reduction of imines with sodium dithionite since the slow reduction was preceded by hydrolysis of the imines. The use of dimethylformamide as a cosolvent with water was reported [8] to increase the reduction rate of aldehydes and ketones. However, in the reduction of imines, the conditions employed by the Dutch group [8] gave products which resulted from hydrolysis of the starting materials.

The method described here overcomes the above difficulties simply by introducing excess sodium dithionite to the solution of imine in hot dimethylformamide, followed immediately by the addition of water. The reaction takes less than 30 min at 110C. Under these conditions, N-cyclohexylidenebenzylamine and N-cyclohexylidenecyclohexylamine give good yields of benzylcyclohexylamine and dicyclohexylamine respectively, while N-benzylidenebenzylamine is converted into dibenzylamine in 55% yield. The more stable and unreactive N-benzylideneaniline is reduced only in moderate yield (40%) to the corresponding amine. While cinnamic acid is unaffected by hot, alkaline sodium dithionite solutions (that is, the conjugated carbon-carbon double bond is not reduced), the reductions of imines of cinnamaldehyde under analogous conditions lead to low yields of complex mixtures.

Lithium aluminium hydride [12] or hydrogen and catalyst [13] reductions of imines containing aromatic carbon-halogen bonds are often accompanied by carbon-halogen bond cleavage; aromatic iodides are particularly susceptible [12]. Sodium dithionite appears to be the reducing agent of choice in such cases. Both iodobenzene and 4-iodobenzoic acid are recovered unchanged after prolonged heating under the normal reducing conditions while N-(3-bromobenzylidene)benzylamine, prepared in situ from 3-bromobenzaldehyde and benzylamine, is reduced cleanly to N-(3-bromobenzyi)benzylamine.

Conversion of Oximes into Corresponding Carbonyl Compounds

When oximes are treated at room temperature with aqueous sodium dithionite, either alone or in the presence of sodium hydrogen carbonate or sodium acetate, the organic material gradually dissolves. The parent carbonyl compound is obtained from this mixture by the addition of acid, preferably, or base.

Two possible mechanisms for the deoximation follow. Firstly, solutions of the dithionite anion are not very stable and decompose in a complex manner to the hydrogen sulfite ion [14]. Hence, the cleavage of oximes with sodium dithionite can occur by a hydrolytic pathway [15] analagous to the reaction of oximes with sodium hydrogen sulfite described by Pine and and coworkers [16]. Alternatively, the cleavage might occur by a reductive pathway, the oxime is first reduced to the imine which is immediately hydrolyzed to the carbonyl compound.

The latter mechanism is supported by the observation that addition of a base apparently increases the rate of the deoximation reaction. Under these conditions, sodium dithionite is known to be a powerful reducing agent [17]. Furthermore, the dithionite deoximations take place under conditions milder than those reported for the hydrogen sulfite reaction [16]. The mechanism is also in keeping with the known reduction of nitro1 and nitroso [5,18] compounds and oximes [5,11] to amines with sodium dithionite and is, furthermore, analogous to the mechanism proposed by Corey and Richman [15] and Timms and Wildsmith [19] for reductive deoximation with Cr(II) and Ti(III) salts respectively.

The dithionite cleavage of oximes competes favourably with known procedures for regeneration of carbonyl compounds from oximes [15,19] The reaction offers the following advantages:

  1. Sodium dithionite is inexpensive and readily available.
  2. The reaction conditions are mild (room temperature at neutral pH).
  3. The regeneration is rapid (several hours at 40C).
  4. Both aldehydes and ketones are regenerated successfully.


Extracts into organic solvents were dried over MgSO4 and evaporated under reduced pressure with a rotary m evaporator. 1H n.m.r. spectra were recorded in CDCl3 on a Varian A60-D spectrometer; chemical shifts were measured from tetramethylsilane as internal standard.

General Procedure for Reduction of Imines

To a solution of the imine (30 mmol), either preformed or prepared in situ from the amine (30 mmol) and the carbonyl compound (30 mmol), in dimethy1formamide (70 ml) at 110C under nitrogen was added solid sodium hydrogen carbonate (120 mmol). The mixture was stirred vigorously; solid sodium dithionite (60 mmol) was added, followed immediately by water (30 ml). Gas evolution took place some minutes after the addition of the water. Stirring was continued at 110C for 30 min; the reaction-mixture was allowed to cool to room temperature and then poured into water (300 ml). The aqueous solution was extracted with ether (4x75 ml) which was in turn washed with water (4x50 ml) and saturated brine (50 ml). The ethereal extract was dried and evaporated to give the amine. The product was purified by distillation or through the hydrochloride.

The amine hydrochlorides were prepared by adding a slight excess of 5 M hydrochloric acid to the neat amine. The mixture was stirred and the solid product was collected by filtration.

  1. N-Cyclohexylidlenebenzylamine (8 g) gave benzylcyclohexylamine (6.8 g, 73%), b.p. 164C/26 mm (lit. [20] 145-147C/115 mm); hydrochloride m.p. 283-284C (sealed tube) (lit. [20] 284C).
  2. N-Cyclohexylidenecyclohexylamine: (10 g) gave dicyclohexylarnine (6.8 g, 68%), b.p. 117-120C/17mm (lit.[21] 118-120C/17 mm); hydrohloride, m.p. 335-337C (lit. 339-342C). (lit.[21] 120C/17mm)
  3. N-Benzylidenebenzylamine (5 g) gave -dibenzylamine (2.7 g, 55%), b.p. 180C/18mm (lit [22] 160-170C/l5 mm); hydrochloride, m.p. 255-260C (lit. [23] 256C).
  4. N-(3-Bromobenzylidene)benzylamine, prepared from 3-bromobenzaldehyde (5 g) and benzylamine (2.9 g) and used crude, gave N-(3-bromobenzyl)benzylamine (3.6 g, 49%), b.p. 220-222C/118mm. The hydrochloride was recrystallized from 0.5 M hydrochloric acid, m.p. 203-205C.
  5. N-Benzylideneaniline (3 g) gave benzylaniline (1.2 g, 40%), m.p. 36-37C, identical with an authentic sample prepared from PhNH2 and PhCH2Cl [24].

Cleavage of Oximes to Carbonyl Compounds

General Procedures

Method A.

The oxime (20 mmol) was mixed with water (15 ml) containing sodium dithionite (28 mmol). The suspension was stirred overnight at room temperature. (Warming to 40C reduced reaction times to several hours.) In some cases, a precipitate formed.This product was very high melting and, on treatment with 2 M hydrochloric acid, liberated the carbonyl compound and sulfur dioxide. It was therefore assumed to be the bisulfite addition compound of the carbonyl compound and was not isolated. A slight excess of 2 M hydrochloric acid was added to the reaction mixture and nitrogen was bubbled through the mixture to expel the sulfur dioxide. Solid sodium carbonate was added carefully to alkalinity; the aqueous mixture was allowed to stand for 30 min and was extracted with ether (2x10 ml) which was dried (MgSO4) and evaporated. The residue was essentially pure carbonyl compound (by t.l.c.)

Method B.

The reaction described under Method A was performed in the presence of sodium hydrogen carbonate (28 mmol). Cleavage by means of this modification appeared to proceed considerably faster. The usual workup gave the carbonyl compound in comparable yield.


  1. Cyclohexanone oxime (2 - 3 g) gradually dissolved in the aqueous sodium dithionite solution when the stirring was continued overnight, a colourless precipitate formed. [This precipitate and, pidly when sodium hydrogen carbonate (Method B) was included in the reaction formed more ra (1.9 g, 95%) was isolated essentially pure. mixture.) Cyclohexanone under the conditions oxime (2 - 7 g reacted sluggishly at room temperature.
  2. Acetophenone rnpleteafter4hat4Oo. Acetophenone (2-2 g,93%) was employed in method A but cleavage was co isolated essentially Pure.
  3. Benzaldehyde oxime(2-4 g) reacted under the conditions of Method A to give a clear solution (At 50, the reaction was complete in 2 h.) Benzaldehyde (2 g, 96%) was isolated essentially pure.
  4. Butanal oxime (2 g) reacted under the conditions of Methods A and B to give a clear solution. Butanal (0- 74 ig) was obtained from the former reaction in 45% yield while, from the latter reaction, butanal was isolated in 54% yield. The relatively low yields were probably due to the high volatility of butanal, b.p. 75C.


[1] J. Prakt. Chem., 1907,76,124.
[2] Ber. Dtsch. Chem. Ges., 1907, 40, 422.
[3] J. Am. Chem. Soc., 1955, 77, 2261.
[4] Chem. Lett., 1977, 1091.
[5] Chem. Ber., 1957, 90, 79.
[6] J. Am. Chem. Soc., 1948, 70, 2656.
[7] Fieser & Fieser, Reagents for Organic Synthesis Vol. 1, p. 1081 (John Wiley, 1968).
[8] Synthesis, 1977, 246.
[9] J. Org. Chem., 1963, 28, 283 1.
[10] J. Org. Chem., 1963, 28, 3259.
[11] J. Am. Chem. Soc. 90, 2927 (1968) and J. Org. Chem. 34, 1817 (1969).
[12] J. Org. Chem. 33, 619 (1968) and J. Org. Chem. 34, 3918 (1969).
[13] J. Org. Chem., 1966, 31, 3875.
[14] Can. J. Chem., 1970, 48, 2778;
[??] Burlamacchi, L., Guarini, G., Trans. Faraday Soc., 1969, 65, 496.
[20] Ber. Dtsch. Chem. G, 1923, 56,1988.
[21] J. Am. Chem. Soc., 1925, 47, 1712.
[22] J. Am. Chem. Soc., 1925, 47, 305 L 1,495, 113.
[23] Kindler, K., Justus Liebigs Ann. Chem., 193
[24] Org. Synth., 1932, Coll. Vol. 1, p 102.