On the Unique Reactivity of Pd(OAc)2 with Organic Azides: Expedient Synthesis of Nitriles and Imines

Organic azides are well-established as versatile compounds that can act as precursors of different heterocycles (triazoles, triazolines, tetrazoles, etc.) or other nitrogen-containing compounds, such as amines (Staudinger reduction, Curtius rearrangement) or imines (Schmidt rearrangement, aza-Wittig reaction).1 Besides the ubiquitous copper-catalysed azide–alkyne cycloaddition reaction,2 two applications of organic azides have recently attracted the interest of the synthetic community: 1) the preparation of aziridines through the generation of nitrenes3 and 2) the synthesis of nitriles. We were particularly interested in the latter application, owing to the importance of the cyano group in industry,4 as well as its utility as an organic synthon. Traditional cyanation methods suffer from serious drawbacks, such as the use of highly toxic reagents, the generation of stoichiometric quantities of metal waste, the need for harsh conditions, and poor functional-group tolerance. 
 
Azides allow the cyanide-free preparation of nitriles with no elongation of the skeletal carbon chain under several conditions: strong stoichiometric oxidants (such as BF3,5a 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ),5b,c or KI/tert-butyl hydroperoxide),5d CuSO4/phenyliodonium diacetate,6 CuI/tert-butyl hydroperoxide,7 or a supported Ru–OH catalyst.8 An earlier report, which employed palladium on charcoal with an alkyne as a hydrogen acceptor, caught our attention for its neutral conditions and its scope, which was not restricted to the formation of benzonitrile derivatives [Eq. (1)].9 
 
 
 
 
 
(1) 
 
 
 
In contrast to the other methods, this Pd system required the rigorous exclusion of moisture and oxygen for any reaction to take place at reflux in benzene or diethylamine. However, despite our efforts, we could not reproduce the reported results and only conversions of <20 % were obtained. Considering that different Pd/C catalysts can lead to adventitious results, we optimised the reaction conditions (Table ​(Table11).10 We only obtained sluggish reactions under an inert atmosphere or in organic solvents (either anhydrous or technical-grade), but the complete conversion of compound 1 a was achieved on water or neat (Table ​(Table1,1, entries 5 and 6). The reactions on water displayed poor selectivity, whereas, in the absence of solvent, anisaldehyde 4 a (15 %) was obtained as the only by-product,11 although it was inseparable from compound 2 a. 
 
 
 
Table 1 
 
Optimisation studies[a] 
 
 
 
If [Pd2(dba)3] was used as the catalyst (Table ​(Table1,1, entry 7), 1,5-diphenyl-3-pentanone was also isolated.12 This compound was generated by the hydrogenation of the dibenzylideneacetone (dba) ligands that were originally on the palladium centre. After screening several Pd0 and PdII sources, Pd(OAc)2 was found to be the best candidate because it suppressed the undesired formation of compound 4 a (imine 3 a was readily separable from compound 2 a) and it was active in technical-grade MeCN at a concentration of only 1 mol % (Table ​(Table1,1, entries 6 and 10). Moreover, we also changed the hydrogen acceptor to styrene, because it is cheaper and both styrene and its reduced form can be easily removed from the crude products under reduced pressure. Notably, azide 1 a was completely converted in the absence of styrene as well, but higher proportions of imine 3 a and the formation of unknown by-products were observed. 
 
Pleasantly, imines 3 were only observed as minor products in the formation of benzonitriles 2 a–2 c and a diverse range of nitriles (2) were isolated in good-to-excellent yields. One exception was phthalimide 2 f, which was isolated together with a bisphthalimide by-product.12 Also, no reaction was observed with cinnamyl azide.13 Terminal alkenes or carboxylic acids did not react under the optimised conditions, despite the generation of H2 (Table ​(Table2,2, entries 9 and 10). The reactions were allowed to proceed for 10 h, but shorter reaction times were sufficient in the cases of, for instance, nitriles 2 g, 2 h, and 2 l.14 
 
 
 
Table 2 
 
Pd(OAc)2-catalysed synthesis of nitriles from azides 
 
 
 
Although organic azides are generally safe and stable towards water and oxygen,1 those of low molecular weight can be particularly dangerous and difficult to handle.15 Gratifyingly, azides 1 could be generated in situ for the one-pot formation of nitriles from their corresponding bromides and NaN3, thereby avoiding the need to isolate any intermediate azides (Table ​(Table3).3). A slightly higher palladium loading (2 mol %) and a concentration of 1 m of the starting bromide maximised the reaction conversion for a range of nitriles that contained different functional groups (Table ​(Table33). 
 
 
 
Table 3 
 
Pd(OAc)2-catalysed synthesis of nitriles from in situ generated azides[a] 
 
 
 
Next, two competition experiments were performed. First, we reacted cinnamyl azide 1 o in the presence of compound 1 g and recovered the former compound unreacted, together with 1-cyanodecane (2 g, Scheme 1 A). All of the previously reported catalytic systems for this transformation converted compound 1 o into its corresponding nitrile.5b,c, 7–9 Furthermore, secondary azides have been reported to produce their corresponding ketones under the same reaction conditions as for the formation of nitriles.11 However, possibly owing to the absence of a strong oxidant, Pd(OAc)2 did not react with secondary azide 1 p, whereas compound 1 l was fully converted (Scheme 1 B). 
 
 
 
Scheme 1 
 
Competition experiments: A) Allyl versus alkyl azides; B) Secondary versus primary azides. 
 
 
 
Regarding the imine formation, Pd(OAc)2 was also the only palladium source that was found to yield compound 3 a as major product if the model reaction was performed under neat conditions.10 This result might not seem surprising, owing to the bimolecular nature of this transformation, but several Pd/C catalysts, as well as [Pd2(dba)3], actually led to the formation of nitrile 1 a as the major product, even in the absence of solvent. Imines are often observed as minor products in the synthesis of nitriles from azides,16 but, to the best of our knowledge, only a molybdenum-based catalyst has been reported to produce them preferentially.17 As a consequence, we prepared several of these derivatives from different benzylic azides.18 In all cases, imines 3 were formed as the major products and they could be separated from their corresponding nitriles (2) by recrystallization or sublimation (Scheme 2). 
 
 
 
Scheme 2 
 
Pd(OAc)2-catalysed synthesis of imines from benzyl azides. 
 
 
 
No styrene was used in the synthesis of imines because it lessened the formation of compound 3 a, thus indicating that dihydrogen might be necessary for the imine formation. Also, nitrile 2 a was integrally recovered after heating for 24 h at 80 °C in the presence of Pd(OAc)2. The reactions shown in Scheme 2 proceeded with evolution of gas within the first five minutes of stirring and the formation of ammonia was indirectly evidenced with wet pH paper. These facts lead us to propose that, in the case of aromatic nitriles, the palladium species in the reaction mixture can hydrogenate compounds 2 to generate a mixture of the imino and amino derivatives that would react together to generate compounds 3 and a molecule of ammonia (Scheme 3).19 Further mechanistic studies on this system are underway. 
 
 
 
Scheme 3 
 
Proposed imine-formation pathway. 
 
 
 
In conclusion, Pd(OAc)2 is an exceptional catalyst for the preparation of nitriles and imines from primary azides (or their corresponding bromides, thus avoiding the need to isolate the intermediary azide). The reactions proceeded under neutral conditions in air and showed unprecedented selectivities in the selected cases shown. The versatility of both azides and nitriles is expected to lead to the widespread application of this system in organic synthesis.


GENERAL CONSIDERATIONS
All reagents, including anhydrous MeCN, were used as received. Unless otherwise stated, all reactions were carried out in air and using technical solvents without any particular precautions to exclude moisture or oxygen. Columns chromatography and TLC were performed on silica gel, using UV light and a phosphomolybdic acid dip to visualize the products. 1 H NMR and 13 C NMR spectra were recorded on 400 MHz spectrometers at room temperature. Chemical shifts (δ) are reported in ppm with respect to tetramethylsilane ( 1 H NMR), or CDCl 3 ( 13 C NMR) as internal standards. Mass spectra (MS) were recorded on a Micromass Autospec Premier, Micromass LCT Premier or a VG Platform II spectometer using EI or ESI techniques at the Mass Spectroscopy Service of Imperial College London.

Methyl 5-azidopentanoate (1k)
Using the general procedure reported by Alvarez

OPTIMISATION STUDIES
All reactions were carried out in technical solvents and in air unless stated otherwise. 1  Spectroscopic data for this ketone were consistent with previously reported data for this compound. 16
Spectroscopic data for 2a were consistent with previously reported data for this compound. 17 -S5 - Spectroscopic data for 1b were consistent with previously reported data for this compound. 17
Spectroscopic data for 2c were consistent with previously reported data for this compound. 18
B: Using the general procedure from 0.137 mL of (2-bromoethyl)benzene, the 1 H NMR spectrum of the crude mixture showed 83% conversion of the starting bromide into the expected nitrile, as well as 12% of azide 1d.
Spectroscopic data for 2d were consistent with previously reported data for this compound. 19
Spectroscopic data for 2e were consistent with previously reported data for this compound. 20

N-(2-Cyanoethyl)-phthalimide (2f)
A: Using the general procedure from 0. Spectroscopic data for 2f were consistent with previously reported data for this compound. 21 N,N'-propylenebisphthalimide was identified by 1 H NMR and mass spectrometry. 22

1-Cyanodecane (2g)
A: Using the general procedure from 0. Spectroscopic data for 2g were consistent with previously reported data for this compound. 21

10-Cyanodecanoic acid (2i)
A: Using the general procedure from 0.227 g of 11-azidoundenoic acid 1i, 0.178 g of the title compound were isolated (90% yield) after purification by flash chromatography on silica gel (EtOAc).
B: Using the general procedure from 0.265 mg of 11-bromoundenoic acid, the 1 H NMR spectrum of the crude mixture showed 81% conversion of the starting bromide into nitrile 2i, as well as 19% of azide 1i.
Spectroscopic data for 2i were consistent with previously reported data for this compound. 23
B: Using the general procedure from 0.219 mL of 11-bromoundec-1-ene, a total conversion into the expected nitrile was confirmed by 1 H NMR. 1 H NMR and mass spectra for 2j were consistent with previously reported data for this compound. 24
Spectroscopic data for 2k were consistent with previously reported data for this compound. 25

SYNTHESIS OF IMINES (3)
C. General procedure for the transformation of azides into imines: In a 20 mL thick-wall scintillation vial, azide 1 (1.0 mmol) and Pd(OAc) 2 (2 mg, 1 mol %) were loaded. The vial was then sealed and heated at 80°C for 10 h. The resulting crude imines were filtered over a plug of celite prior purification. The reported yields are isolated yields are the average of at least two independent runs.
-S9 -Spectroscopic data for 3a were consistent with previously reported data for this compound. 26
Spectroscopic data for 3b were consistent with previously reported data for this compound. 26
Spectroscopic data for 3x were consistent with previously reported data for this compound. 27
Spectroscopic data for 3m were consistent with previously reported data for this compound. 28

N-Benzylidenebenzylamine (3n)
C: Using the general procedure from 125 µL of benzyl azide 1n, 74 mg of the title compound were isolated (76% yield) after concentration under high vacuum.
Spectroscopic data for 3n were consistent with previously reported data for this compound. 26 -S10 -