Significant interest has developed in recent years in the application of iminophosphoranes for various synthetically useful transformations, especially those affording C=N bond possessing compounds, which are usually referred to as aza-Wittig reactions¹. Although numerous articles have appeared on the reactions and synthetic applications of N-alkyl(aryl)triphenyliminophosphoranes², the preparative utility of analogous aza-ylides, viz. N-allyltriethoxyiminophosphoranes is almost unexplored³. Recently, we reported the one-pot preparation of N-alkyltriethoxyiminophosphoranes from alkyl bromides by azidation followed by Staudinger reaction⁴ and the application of these compounds for the synthesis of primary amines⁵. Herein we describe a route for the synthesis of secondary amines beginning with the parent alkyl bromides and employing a facile construction of the final carbon skeleton via the aza-Wittig reaction between N-alkyltriethoxyiminophosphoranes and aldehydes. Secondary amines are synthesized by the variety of methods, the reduction of imines being a very convenient and explicit route⁶. Difficulties which may be sometimes encountered in this approach are due to the properties of some imines which are unstable and therefore not easily accessible in pure state. We decided to overcome this disadvantage by preparing the respective imines under mild conditions and use them for reduction without isolation and purification. The synthetic protocol illustrating the construction of secondary amine skeleton from two simple building blocks - alkyl bromide and aldehyde is presented in the Scheme.

Azidation of alkyl bromides was performed for 3 hours at the reflux temperature of the solvent system, using a 100% excess of sodium azide. Benzene solution of crude alkyl azides 1 afforded N-alkyl- triethoxyiminophosphoranes 2 when treated with triethyl phosphite at 25-30°C.

Isolation and purification of these aza-Wittig reagents was neither necessary nor recommended. Neat 2 are extremely sensitive to traces of moisture and hydrolyze easily to diethyl N-alkylphosphoramidates which may then contaminate the target compounds. Benzene solution of 2 reacted easily with aliphatic and aromatic aldehydes to give the imines 3. If one followed the course of this reaction by ³¹P-NMR, the practically complete conversion of 2 to 3 was observed after 3 hours at room temperature. Due to the exothermic nature of this reaction, all reactions were started at 0-5°C and then run at 20-25°C.

The crude imines 3 were not isolated but immediately reduced with sodium borohydride in methanol to form the target secondary amines 4. The amines were purified by steam distillation and characterized as corresponding hydrochlorides 5. This methodology can be used to prepare a number of secondary amines as shown in Table 1. The range of substrates in Table 1 demonstrates the clear utility of this synthetic procedure in preparing a variety of secondary amines in moderate to good yields. Best results were obtained when benzaldehyde and primary alkyl bromides were used as starting materials. The overall yields of 5 were inevitably lower (entry 6 and 7) with easily enolizable butyraldehyde as carbonyl substrate, possibly due to base-promoted (by strongly basic 2) aldol condensation as a side reaction.

Although secondary alkyl bromides could be easily converted into the corresponding iminophosphoranes 24, the aza-Wittig reactions of the latter were totally ineffective (entry 8).

It is well known that reaction of azides with tertiary phosphines gives iminophosphoranes¹, useful intermediates for the synthesis of a variety of nitrogen compounds² including imines³. As part of our program on the photochemistry of some allyl imines, we directed our efforts toward their efficient preparation. Now we wish to report an effective one-pot synthesis of imines directly from alkyl bromides via azide intermediates under the conditions of the Staudinger reaction in acetonitrile.

The literature reports a great number of procedures for the synthesis of alkyl azides by the nucleophilic substitution of bromide with sodium azide⁴. To obtain a corresponding imine, isolated alkyl azides are generally treated with triphenylphosphine and a carbonyl compound in a nonpolar solvent such as benzene³. Those procedures are usually too complex and so we decided to find a short, simple, and efficient method.

Acetonitrile was found to be a very practical solvent in both steps: azidation and imine formation. Thus, both steps were combined without necessity to isolate and purify any intermediate (Scheme 1). Reactive primary bromides were refluxed with an excess of sodium azide in acetonitrile. Sodium bromide and the remaining sodium azide were easily filtered off from the cooled reaction mixture because both salts are insoluble in cold acetonitrile. Since aliphatic nonyl bromide did not afford nonyl azide under those conditions, a small amount of water or a phase-transfer catalyst were used to enhance the reactivity. The presence of a phase-transfer catalyst raised the conversion, however, the reaction time was too long (over 50 hours). Reflux of the acetonitrile solution with triphenylphosphine and an aldehyde afforded the corresponding imine in a very high yield and purity (Table 1). Reaction of cinnamyl bromide provided imine with only 70% purity - some unidentified byproducts were formed. When ketone instead of aldehyde was used a longer reaction time in the second step was applied bat the conversion was still too low.

The described one-pot procedure offers relatively high overall yields and thus may serve as an helpful sequence in many procedures whom the synthesis of imines is anticipated.

The conversion of aldehydes and ketones into imines and the subsequent synthesis of α-unsubstituted and a-branched secondary amines by the reduction of the imine,^1,2 by the 1,2-addition of organometallic reagents to the imine,^3,4 or by the addition of other nucleophiles to the imine⁵ is of current and continuing importance. Several recent reports have been concerned with the use of solid phase methods for achieving such transformations.^6-8 The use of azides as the nitrogen source for the synthesis of primary amines can be achieved by the reduction of the azide or by the conversion of the azide into an iminophosphorane [the Staudinger reaction] followed by hydrolysis,⁹ a process which has been transferred to the solid phase.¹⁰ These methodologies do not involve imine intermediates and cannot be used to access secondary amines. The solution phase conversion of azides into Boc-protected amines has been reported,¹¹ and other methods for the conversion of azides into secondary amines which do not involve iminophosphorane or imine intermediates are also known.¹² We are aware of only a few isolated specific examples for the solution phase conversion of azides into secondary amines,^13a particularly polyamines,^13b via intermediate iminophosphoranes and imines. However, these reports do not disclose widespread applicability and no detailed or general studies have appeared to date. No solid phase applications for the conversion of azides into α-unsubstituted or α-branched secondary and primary amines via intermediary imines and iminophosphoranes have been reported.¹⁴

In this preliminary report, we present a simple, general, one-pot solution phase and polymer supported protocol for the combination of azide and aldehyde starting materials which allows access to α-unsubstituted and α-branched secondary and primary amines. The chemistry proceeds via the in situ reactions of intermediate iminophosphoranes and imines, where the iminophosphoranes may be polymer supported. The solid phase process offers significant advantages over the corresponding solution phase method in terms of time, ease of purification, avoidance of chromatography and cost. A clear advantage of the solid support methodology is its potential for the generation of combinatorial libraries of amines using readily available starting materials which offer several points of diversity.

We began our study by establishing the validity of the chemistry in the solution phase, as shown in Scheme 1. Thus, the Staudinger reaction¹⁵ between the phosphine 2 and the azide 1 gave the iminophosphoranes 3, which were then made to undergo in situ aza-Wittig reaction¹⁶ with aldehydes 4 to give the imines 5. The in situ reduction of the imines 5 with sodium borohydride, sodium cyanoborohydride or borane-THF gave, after flash silica chromatography, the α-unsubstituted amines 6, generally in high yields as shown in the Table. The in situ reaction of the imines 5 with an organometallic (R₃Li or R₃MgX) similarly allowed access to the α-branched amines 7, again in high yields. The range of suitable organometallic reagents included allyl magnesium bromide, thereby allowing access to synthetically useful homoallylic amines.^4a,7a,17 The use of benzyl azide allowed a practical synthesis of N-benzyl protected amines whilst the use of trimethylsilyl azide (1, R = TMS) allowed access to primary amines 8, a process which proceeds via the in situ N-desilylation¹⁷ of the intermediate N-silylated amines^4a,17 upon workup.

The Table shows the azides, aldehydes and organometallics that were used in our study, together with the isolated yields of the amines. In the solution phase, we have also screened the use of a range of trivalent phosphorus species 2. We found that the use of tri(n-butyl)phosphine typically gave yields of amines 5-10% higher than triphenylphosphine. The use of trimethyl phosphite, triethyl phosphite and hexamethylphosphorous triamide allowed the process to be carried out at lower temperatures than with trialkyl- or triaryl-phosphines without significant differences in yield. However, we believe that ease of handling, odour containment and other safety issues make triphenylphosphine the phosphine of choice for this process, and the Table refers to results obtained with this phosphine.¹⁸

Gratifyingly, all stages of this chemistry were compatible with transfer to a polymer supported protocol by the use of polymer supported triphenylphosphine¹⁹ at the phosphine stage to give the polymer supported iminophosphorane 10 as shown in Scheme 2.²⁰ Advantageously, we found that polymer supported cyanoborohydride brought about the reduction of the intermediate imines 5 in good yields, as shown in the Table.^20,21 Chromatographic separation of the phosphine oxide was not required: the polymer supported phosphine oxide 11 was removed by filtration. Polymer supported phosphine 9 can be regenerated by the reduction of polymer supported phosphine oxide 11 with trichlorosilane.^19a The use of Amberlite® azide exchange resins for the synthesis of azides and the availability of polymer supported methods for the synthesis of aldehydes from alcohols^8a offer further possibilities which may be significant for future applications in combinatorial chemistry.

In conclusion, we have developed a simple polymer supported method for the combination of azides with aldehydes to give α-unsubstituted and α-branched primary and secondary amines. Further studies on the development of methodologies for the polymer supported asymmetric synthesis of α-branched amines and the polymer supported synthesis of nitrogen containing heterocycles based upon Staudinger/aza-Wittig chemistry are underway.