First generation of pentazole (HN₅, pentazolic acid), the final azole, and a zinc pentazolate salt in solution: A new N-dearylation of 1-(p-methoxyphenyl) pyrazoles, a 2-(p-methoxyphenyl) tetrazole and application of the methodology to 1-(p-methoxyphenyl) pentazole

Abstract

Ceric ammonium nitrate (CAN) in methanol–water gave a new N-dearylation of a series of substituted 1-(p-methoxyphenyl) pyrazoles and a 2-(p-methoxyphenyl)tetrazole producing p-benzoquinone and the parent azole in a mole for mole ratio. Application of this reaction to 1-(p-methoxyphenyl) pentazole at −40 °C produced p-benzoquinone. ¹⁵N NMR spectra suggest that pentazole, HN₅, was also produced and held in solution as N₅⁻ with Zn²⁺ ion. The ¹⁵N signal from N₅⁻ was −10.0 ± 2.0 ppm in agreement with calculated values.

---

The existence of an all-nitrogen aromatic azole ring, R–N₅, the pentazoles, was proved by Clusius and Hurzeler¹ in Zurich and Huisgen and Ugi² in Munich.³ Since then a significant goal of heterocyclic chemistry has been to prepare the parent pentazole, HN₅. Ozonolytic degradation of the aryl ring was attempted⁴ and recently revisited with controlled limited amounts of ozone,⁵ but HN₅ or its anion N₅⁻ was not found. Following the recent discovery of N₅⁺ as a reasonably stable cation in inorganic salts^6,7 there has been renewed experimental interest in allotropes of nitrogen and in N₅⁻ in particular. There has also been much theoretical interest in these types of species.^8–12 Metallocene pentazole derivatives have been predicted as viable forms of N₅⁻, but bidentate forms MN₅ (M = Na, K, MgCl⁺, CaCl⁺) and M(N₅)₂ (M = Na, K, Mg, Ca) and unidentate Zn(N₅)Cl and Zn(N₅)₂ are calculated to have lower energy.¹³ Recently it has been suggested the N₅⁺.N₅⁻ is a viable entity.¹⁴

Following our recent establishment of the mechanism of formation of aryl pentazoles from aryldiazonium ions and azide ion,¹⁵ we have been endeavouring to generate HN₅ and its anion N₅⁻. Recently N₅⁻ has been detected for the first time from high voltage collisions of 4-pentazolylphenolate ion, N₅-C₆H₄O⁻, in the mass spectrometer.¹⁶ We are therefore prompted to make a preliminary report of our work.

To remove an aryl group from an arylpentazole without destruction of the pentazole ring a mild oxidizing agent and the appropriate aryl pentazole must be chosen. Ceric (IV) ammonium nitrate (CAN) has been used to remove p-MeOC₆H₅- groups from p-methoxyphenyl ethers^17,18 and N-p-methoxyphenyl azetidines.^19–20 The methoxyphenyl ring is removed as p-benzoquinone and the use of H₂¹⁸O has proved that the two quinone oxygens arise from water in the solvent.¹⁸ Successive one-electron oxidations are considered to produce a quinone-imine dication which is hydrolysed.²⁰ We tested this reaction for the first time with an azole substrate using the 1-N-p-methoxyphenyl pyrazole series shown in Fig. 1. It proved successful and was optimised at low temperatures. The p-benzoquinone is easily detected and estimated. Its presence indicates a mole for mole presence of the azole. High temperatures are not appropriate for this reaction because the azole generated may undergo addition reactions with p-benzoquinone.²¹ We also attempted the reaction with the non-metallic one-electron acceptor tetracycanoethylene,²² but this gave no dearylation. A similar CAN oxidation of the tetrazole 4 gave p-benzoquinone and 5-phenyltetrazole 5.

Fig. 1 ^a Isolated yield (%) at ambient temperature. ^b Isolated yield (%) at –10 °C. Reagents: (i) Ce(NH₄)₂(NO₃)₆ in MeCN∶H₂O, (83∶17 v/v) or MeOH∶H₂O (80∶20 v/v).

A theoretical study was carried out¹³ to determine the ¹⁵N NMR shift expected for N₅⁻ and to identify the best cation to retain it, which proved to be Zn²⁺. An analysis of LCAO-MO coefficients showed considerable interaction between Zn d orbitals and the N₅⁻ lone pairs and π MOs.¹³ A Mulliken population analysis indicated much more of a covalent interaction between Zn and N₅⁻ than the other metals. The Mulliken charges on the metal calculated with the basis set used in our study are Na +0.805, K +0.914 for MN₅, and Mg +0.756, Ca +0.869, Zn +0.300 for M(N₅)Cl. Kaszynski et al.⁵ have concluded that HN₅ has a half-life of only 10 min and hence it could probably not be directly detected. It has also been suggested that HN₅ is a stronger acid than nitric acid,²³ hence if generated a significant portion would be converted to the anion and held were an appropriate cation such as Zn²⁺ present.

When the CAN dearylation was applied to 1-(p-methoxyphenyl) pentazole at −40 °C, p-benzoquinone was readily detected in the solution by tlc and confirmed by direct detection with ¹H and ¹³C NMR spectra. Proton NMR and GC calibration against standard benzoquinone solutions gave yields of up to 25% quinone. The benzoquinone was also isolated as yellow crystals by sublimation. The presence of the benzoquinone suggests that HN₅ has been generated. The pentazole used was prepared as described by Huisgen and Ugi.² When NMR spectra were measured however there was always some p-methoxyphenyl azide present but ratios of pentazole to azide as high as 11∶1 were achieved. Fig. 3a shows the ¹⁵N NMR spectrum of 6 where signals from the pentazole only are seen. The proton NMR spectrum of the same solution (Fig. 3a, inset) shows the presence of 7% p-methoxyphenyl azide. Control reactions of CAN with p-methoxyphenyl azide produced effectively no measurable p-benzoquinone and the small impurity of aryl azide could not account for the observed results from the reaction of CAN with 6.

Fig. 2 ^a15N shifts ppm from MeNO₂ in CD₃OD. ^b15N shifts ppm in CD₃OD∶D₂O 77∶23 (v/v). Reagents: (i) Ce(NH₄)₂(NO₃)₆ (2.8 mol), Zn(NO₃)₂·6H₂O (6 mol), CD₃OD∶D₂O 77∶23 (v/v).

Fig. 3 (a) ¹⁵N NMR spectrum of ¹⁵N-labelled 6 (see Fig. 2) with inset ¹H NMR spectrum showing 7% arylazide (cf. ref. 24). (b) ¹⁵N NMR spectrum of reaction mixture of CAN and compound 6 when held at −40 °C for up to 2 weeks. (c) ¹⁵N NMR spectrum of reaction mixture after 7 days at −40 °C followed by 68 h at −20 °C, showing ¹⁵N₂ (−72 ppm)²⁴ and ¹⁵N₃⁻(−283 ppm, −147 ppm). Larger version of spectra available in ESI.†

The key experiments shown in Fig. 2 were then performed.† The ¹⁵N-labelled derivative 6 was prepared from the reaction of p-methoxyphenyl diazonium ion with terminal ¹⁵N-labelled azide ion which places ¹⁵N atoms at the 2-,3-,4- and 5-pentazole positions.¹⁵ Solutions of 6 in CD₃OD:D₂O containing excess Zn(NO₃)₂ were treated with CAN at −40 °C and followed by ¹⁵N NMR spectra.† Only ¹⁵N labelled nitrogen atoms are detected. As the signals of the pentazole declined a new signal appeared at −10.0 ± 2.0 ppm in different runs (Fig. 3b). The calculated ¹⁵N shift for Zn(N₅)Cl is −16.6ppm.¹³ This is an average of the five values reported for the ¹⁵N shifts. Geometry optimization of Zn(N₅)NO₃ using the same procedure¹³ leads to the unidentate, planar configuration and an average ¹⁵N shift of −17.5 ppm. The spread in the signal position, −10.0 + 2.0 ppm in different runs, may be due to the mix of ions present in the solution. The cations present are Zn²⁺, NH₄⁺, Ce(IV) and Ce(III) and the anions are NO₃⁻ and N₅⁻. Ce(III) is expected to be detrimental to N₅⁻. When the solution was warmed to −20 °C for ca. 24 h a signal at −72 ppm due to ¹⁵N-labelled nitrogen gas appeared along with signals at −283 ppm and −147 ppm, the terminal and central atoms of ¹⁵N₃⁻ respectively (Fig. 3c). When the solution was warmed above 0 °C these three signals disappeared. At the end of the reaction there were no signals for ¹⁵N-labelled azide ion and careful work-up of aqueous and organic extracts showed no traces of N₃⁻ ion in the IR spectra of the residue. The anion N₅⁻ would be expected to break down to ¹⁵N₂ and ¹⁵N₃⁻ (as observed herein and also in gas phase¹⁶) since ArN₅ breaks down to ArN₃ and N₂. We had expected the ¹⁵N₃⁻ to grow and remain in the solutions. Optimising the geometries of the reactants and products at the B3LYP/6-11++Gdp level of theory¹³ gives a reaction energy of

N₅⁻ → N₂ + N₃⁻

ΔH = −59 kJ mol⁻¹

(−60 kJ mol⁻¹ at the CCSD(T)/aug-cc-pVTZ level of theory¹⁴).

However the presence of excess NH₄⁺ ions in the solutions can give rise to the following highly exothermic removal of azide ion:

N₅⁻ + 3 NH₄⁺ + 2 N₃⁻ → 5 N₂ + 4 NH₃

ΔH = −2558 kJ mol⁻¹

Even when considering the effect of counter ions as ion pair structures the reaction is highly exothermic:

Zn(N₅)NO₃ + 3 NH₄NO₃ + 2 Zn(N₃)NO₃ → 3 Zn(NO₃)₂ + 5 N₂ + 4 NH₃

ΔH = −2537 kJ mol⁻¹

The totality of the experimental results we have observed suggest that in these solutions we have generated HN₅ and held it for a time in a zinc(II) salt.

From literature work^17–20 we believe the key intermediate in these dearylation reactions is the species 8. This must not be looked upon as an unstable organic dication. When two NO₃⁻ ions pairs above and below the benzene ring of species 8 are included in B3LYP/3-611 + G(d) geometry optimisations (cf. ref. 13 for details of the method), considerable stabilisation of the organic component is found. Bond lengths resemble much more a neutral aromatic system than a quinone structure. The 1,4-benzene carbons remain positive enough to favour nucleophilic H₂O addition which leads to the separation of the two aromatic rings.

Notes and references

1. K. Clusius and H. Hurzeler, Helv. Chim. Acta, 1954, 37, 798.
2. R. Huisgen and I. Ugi, Chem. Ber., 1957, 90, 2914; R. Huisgen and I. Ugi, Angew. Chem., 1956, 68, 705.
3. A review, R. N. Butler, Comprehensive Heterocyclic Chemistry II, Series eds., A. R. Katritzky, C. W. Rees, E. F. V. Scriven, Pergamon Press, Oxford, 1996, vol. 4 (ed. R. C. Storr), p. 897.
4. I. Ugi, Angew. Chem., 1961, 73, 172.
5. V. Benin, P. Kaszynski and J. G. Radziszewski, J. Org. Chem., 2002, 67, 1354.
6. K. O. Christe, W. W. Wilson, J. A. Sheehy and J. A. Boatz, Angew. Chem., Int. Ed., 1999, 38, 2004.
7. A. Vij, W. W. Wilson, V. Vij, F. S. Tham, J. A. Sheehy and K. O. Christe, J. Am. Chem. Soc., 2001, 123, 6308.
8. M. T. Nguyen, M. Sana, G. Leroy and J. Elguero, Can. J. Chem., 1983, 61, 1435.
9. M. T. Nguyen, M. A. McGinn, A. F. Hegarty and J. Elguero, Polyhedron, 1985, 4, 1721.
10. R. J. Bartlett, Chem. Ind. (London), 2000, 140; K. F. Ferris and R. J. Bartlett, J. Am. Chem. Soc., 1992, 114, 8302.
11. G. A. Olah, G. K. Surya Prakash and G. Rasul, J. Am. Chem. Soc., 2001, 123, 3308.
12. A. Hammerl and T. M. Klapötke, Inorg. Chem., 2002, 41, 906.
13. L. A. Burke, R. N. Butler and J. C. Stephens, J. Chem. Soc., Perkin Trans. 2, 2001, 1679.
14. M. T. Nguyen and T. K. Ha, Chem. Phys. Lett., 2001, 335, 311; S. Fau, K. J. Wilson and R. J. Bartlett, J. Phys. Chem. A, 2002, 106, 4639.
15. R. N. Butler, A. Fox, S. Collier and L. A. Burke, J. Chem. Soc., Perkin Trans. 2, 1998, 2243; R. N. Butler, S. Collier and A. Fleming, J. Chem. Soc., Perkin Trans. 2, 1996, 801.
16. A. Vij, J. G. Pavlovich, W. W. Wilson and K. O. Christe, Angew. Chem., Int. Ed., 2002, 41, 3051 , cf. Chem. Eng. News, August 19, 2002, p.8.
17. G. A. Molander, Chem. Rev., 1992, 92, 29.
18. P. Jacob III, P. S. Callery, A. T. Shulgin and N. Castagnoli Jr., J. Org. Chem., 1976, 41, 3627.
19. D. R. Kronenthal, C. Y. Han and M. K. Taylor, J. Org. Chem., 1982, 47, 2765.
20. J. Fetter, E. Keskeny, T. Czuppon, K. Lempert, M. Kajtár-Perody and J. Tamás, J. Chem. Soc., Perkin Trans. 1, 1992, 3061.
21. P. Ballesteros, R. M. Claramunt, C. Escolástico, M. D. Santa Maria and J. Elguero, J. Org. Chem., 1992, 57, 1873.
22. R. N. Butler, J. Oakes and M. C. R. Symons, J. Chem. Soc. Section A, 1968, 1134.
23. C. Cheng, Int. J. Quantum Chem., 2000, 80, 27.
24. R. Müller, J. D. Wallis and W. von Philipsborn, Angew. Chem., Int. Ed., 1985, 24, 513.

---

Footnote

† Electronic supplementary information (ESI) available: experimental details. See http://www.rsc.org/suppdata/cc/b3/b301491f/

---