Nucleophilic Attack of Azide at Electrophilic Azides: Formation of N₆ Units in Hexazene and Aminopentazole Derivatives

Abstract

With the help of selective ¹⁵N labeling experiments, it has been confirmed that nucleophilic attack of azide at iminium‐activated organic azides leads to short‐lived hexazene intermediates. Such species do not only tend to a cleavage reaction with formation of N‐azido compounds, but also undergo ring closure to generate unprecedented amidino‐functionalized pentazoles. Thus, treatment of the parent Vilsmeier reagent with two equivalents of sodium azide creates an aminopentazole derivative as the main product, which is easily characterized by NMR spectroscopy.

Simple substrates, all commercially available, lead to previously unknown aminopentazole derivatives. By using Na¹⁵N₃, the heterocyclic product with six‐fold ¹⁵N labelling is easily accessible, which enabled structure identification by ¹⁵N NMR spectroscopy without comparison with known data.

Initial attempts to generate an all‐nitrogen five‐membered ring were performed more than a hundred years ago.1 But the first report with evidence of the arylpentazole 2 dates back from the year 1956 when Huisgen and Ugi described experiments with such intermediate species in the transformation of aryldiazonium salts 1 to prepare aryl azides 3 (Scheme 1 a).2 Later, isolation and even characterization of pentazole 2 (Ar=4‐Me₂NC₆H₄) with the help of single‐crystal X‐ray diffraction analysis were successful.3 However, attempted modification of the aryl group of 2 have usually resulted in destruction of the pentazole ring, which degraded rapidly at ambient temperature with evolution of dinitrogen.4 To our knowledge, the reaction of substrates 1 with azide salts proves to be the only method for the synthesis of pentazoles, and aryl derivatives of type 2 are the only representatives which could be prepared so far.5

Scheme 1.

Synthesis of penzatoles 2 and 7 as well as pentazolate salts 5.

After a long period of unsuccessful attempts and hundreds of experiments, reaction conditions were recently found for the cleavage of arylpentazole 2 (Ar=3,5‐dimethyl‐4‐hydroxyphenyl) in the presence of m‐chloroperbenzoic acid (m‐CPBA) and ferrous bisglycinate [Fe(Gly)₂].6 This transformation led to the salt 5 (M=NH₄) and effected a breakthrough in pentazolate chemistry and also in the synthesis of other salts of type 5.7 Nearly simultaneously, the products 5 (M=Cs, Li) were generated by laser heating of alkali azides 4 in the presence of dinitrogen under very high pressure (Scheme 1 a).8 Currently, polynitrogen compounds, such as pentazoles and pentazolate salts, attract attention because they are assumed to have important applications as high energy density materials (HEDMs).9

Herein, we report an unprecedented synthesis of the aminopentazole derivative 7, which is available by treating the commercial Vilsmeier reagent 6 with sodium azide (Scheme 1 b). The product 7 is probably formed via cyclization of a short‐lived hexazene derivative that is generated by nucleophilic attack of azide at an iminium‐activated organic azide. This assumption is based on mechanistic studies with other chloroiminium salts (for example, see 8) exposed to sodium azide.

Our approach to prepare aminopentazole derivatives was based on the work of Balli and co‐workers,10 who synthesized the 2‐azidobenzothiazolium salt 9 by treating the precursor 8 with one equivalent of sodium azide (Scheme 2). The reaction of 9 with lithium azide in dimethylformamide led to the highly unstable N‐azido compound 13. The salt 10, which should establish an equilibrium with the covalent diazide 11,11 and the hexazene12 derivative 12 were postulated as short‐lived intermediates to explain the formation of the unusual final product 13.10

Scheme 2.

Generation of the N‐azido compound 13.

Quite recently, it was shown that the structures of several N‐azidoamines, which were previously reported in the literature, are not correct and have been revised.13 Thus, we assumed that structural verification of the N‐azido compound 13 might be useful since the former characterization was mainly based on IR data10 owing to the low stability of this substance. The reaction mechanism including nucleophilic attack of azide at the terminal nitrogen atom of the iminium‐activated azido group of 10 and formation of intermediate 12 is plausible but also an unprecedented case. We thought that migration of an azido group of diazide 11 accompanied by liberation of dinitrogen may alternatively lead directly to 13 without the creation of 12. If hexazene intermediate 12 is really generated, however, cyclization to produce the corresponding aminopentazole derivative is possibly observable at low temperature. This assumption is based on the well‐known ring closure of pentazenes to yield pentazoles.5, 14

When we treated the substrate 9 in d₇‐DMF with hexadecyltributylphosphonium azide (Scheme 3), which includes a ¹⁵N label at one of the terminal nitrogen atoms (98 %),15 we obtained the desired N‐azido compound ¹⁵N₃‐13 quantitatively (¹H NMR spectroscopy at −60 °C). The ¹⁵N NMR spectrum of this product indicated that the label was distributed among the imino group as well as N‐β and N‐γ of the azido group. Both signals of the azido group were accompanied by small doublets, which revealed a direct coupling of the two nitrogen atoms with ¹ J(¹⁵N,¹⁵N)=9.3 Hz. These results show that the equilibration of isotopically labeled intermediates 10 and 11 led to such species with two, one, or zero ¹⁵N atoms. The distribution of the ¹⁵N label, observed in ¹⁵N₃‐13, is only compatible with formation of 13 via cleavage of 12.16 In the alternative case with liberation of dinitrogen from 11, the isotopically isomeric product ¹⁵N₃‐13′ should be generated. We treated 13 and ¹⁵N₃‐13 with cyclooctyne17 in order to confirm the N‐azido structure and the distribution of the ¹⁵N label. The product ¹⁵N₃‐14 was formed quantitatively (¹H NMR spectroscopy) and led to ¹⁵N NMR signals at δ=−158.7, −48.7, and −44.7 ppm, which were compared with the ¹⁵N NMR spectrum of 14 measured with natural abundance (Scheme 3). The signals of ¹⁵N₃‐14 with δ=−48.7 and −44.7 ppm were accompanied by small doublets with ¹ J(¹⁵N,¹⁵N)=18 Hz. Whereas the 2D‐¹⁵N,¹H shift correlation spectrum of 14 showed cross signals for N‐3′/4′‐H as well as N‐1′/9′‐H, the corresponding spectrum of ¹⁵N₃‐14 indicated the former cross signal only, and a correlation of the 9′‐H signal with nitrogen signals was not observed.

Scheme 3.

Synthesis of the ¹⁵N‐labeled compounds 13 and 14 as well as ¹⁵N NMR chemical shifts (δ) of 14.

Our investigations allow the drawing of the following interim conclusions: The ¹⁵N NMR data and the trapping reaction of 13 with the help of cyclooctyne confirm the N‐azido structure of 13. Moreover, it is demonstrated that nucleophilic attack of azide at iminium‐activated organic azides is possible since 13 was generated from 9 via hexazene derivative 12. Disappointingly, cyclization of short‐lived 12 to create an aminopentazole derivative was not observed.

When we planned and performed final experiments to additionally verify the ¹⁵N NMR data of 13 by synthesizing ¹⁵N₄‐13 from precursor 8 and two equivalents of fully labeled sodium azide (Na¹⁵N₃, 98 %), we obtained not only the desired product ¹⁵N₄‐13, but also the surprising compounds 15, ¹⁵N₆‐7‐d₇, and ¹⁵N₄‐17‐d₇ (Scheme 4). Clearly, the formation of the unexpected products was connected with a halogen/oxygen exchange reaction of the solvent d₇‐dimethylformamide and the substrate 8 to form the deuterated Vilsmeier reagent 6‐d₇.18 We assumed that 6‐d₇ was transformed into the hexazene derivative ¹⁵N₆‐16‐d₇ by double nucleophilic attack of azide, which is similar to the creation of 12 from 8. Whereas 12 exclusively underwent liberation of dinitrogen to produce 13, the short‐lived intermediate ¹⁵N₆‐16‐d₇ preferred cyclization leading to the aminopentazole derivative ¹⁵N₆‐7‐d₇, and the N‐azido compound ¹⁵N₄‐17‐d₇ was formed as a minor product. When we analogously treated Vilsmeier reagent 6 with fully ¹⁵N labeled sodium azide or with selectively labeled Na¹⁵N=N=N, we obtained ¹⁵N₆‐7 and ¹⁵N₄‐17 or ¹⁵N₄‐7 and ¹⁵N₃‐17, respectively. If d₇‐dimethylformamide was used as solvent for the reaction of 6 with fully ¹⁵N labeled sodium azide, the halogen/oxygen exchange reaction of the deuterated solvent and 6 caused generation of 6‐d₇ and thus creation of ¹⁵N₆‐7‐d₇ and ¹⁵N₄‐17‐d₇. In all cases, the molar ratio of 7 to 17 was approximately 2:1.

Scheme 4.

Synthesis of aminopentazole derivatives 7 from Vilsmeier reagent 6 and sodium azide.

The N‐azido compound 17 can easily be handled in solution at −40 °C; however, rapid decay with a half‐life t _1/2 of approximately 16 min, which was measured by collecting the liberated dinitrogen gas, was observed at −30 °C.16 Hence, 17 is significantly less stable than 13, and consequently, 17 did not undergo a clean trapping reaction with cyclooctyne.19 On the other hand, solutions of the aminopentazole derivative 7 can be utilized for NMR spectroscopy at +10 °C, and a half‐life t _1/2 of around 11 min was roughly estimated at 21 °C. The identification of 7 and 17 was mainly based on NMR spectroscopy and especially ¹⁵N NMR data. The ¹⁵N NMR data of ¹⁵N₄‐17 and those of ¹⁵N₄‐13 are very similar.16 The ¹⁵N NMR chemical shifts and the ¹⁵N,¹⁵N coupling constants of ¹⁵N₆‐7 and ¹⁵N₄‐7 are in excellent agreement with those published for several other pentazoles.4b, 5a, 20 To our knowledge, ¹⁵N₆‐7 is the first pentazole derivative in which all members of the ring are labeled by ¹⁵N atoms. Therefore, there is no need to compare with known data of other pentazoles since the coupling patterns alone are unequivocal proof of the structure with an all‐nitrogen five‐membered ring connected with a sixth nitrogen atom. As depicted in Figure 1, the imine nitrogen atom couples with N‐1 (¹ J=−14.3 Hz), and a direct coupling with N‐2 and N‐5 is responsible for the additional triplet splitting of the N‐1 signal (¹ J=−18.0 Hz). Other triplet splittings were detected for the N‐1 signal by geminal coupling with N‐3 and N‐4 (² J=0.8 Hz) and also for the imine signal by geminal coupling with N‐2 and N‐5 (² J=2.2 Hz). Finally, N‐2, N‐5, N‐3, and N‐4 create an AA′XX′ system, which was analyzed by iterative simulation of the ¹⁵N NMR spectrum.16

Figure 1.

¹⁵N NMR spectrum of ¹⁵N₆‐7 and ¹⁵N₄‐17 in d₇‐DMF measured at −60 °C (61 MHz, reference MeNO₂ with δ=0; J values in Hz).

In conclusion, we confirmed the nucleophilic attack of azide at the terminal nitrogen atom of iminium‐activated organic azides. Clearly, the resulting short‐lived hexazene derivatives undergo not only a cleavage reaction to generate N‐azido compounds, but also a cyclization leading to unprecedented aminopentazole21 structures. This simple access to amidino‐substituted pentazoles is remarkable, especially as the precursor 6 is a commercial substance known as the Vilsmeier reagent and isolated in 1959 for the first time.22 Currently, we investigate whether the new approach to functionalized pentazoles can be transferred to other chloroiminium substrates. Preliminary experiments have shown that pentazolate salts are also available by similar reactions. Moreover, we assume that the decay of 7 and 17 will offer an access to dimethylaminomethylidene, a rarely studied carbene.23 This expectation is based on the known decomposition reaction of 13, which led to the corresponding short‐lived benzothiazol‐2‐ylidene.10g

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

K. Banert, T. Pester, Angew. Chem. Int. Ed. 2020, 59, 12315.