We initiated a study of the possible thermolytic generation and reaction of the 4-(1,2,4-triazolyl) cation (1) from readily available and inexpensive 4-amino-1,2,4-triazole,^{6, 9} and have extended it to the photolysis of suitable nitrenium ion precursors. We now describe these studies in full.

    We have shown that N-amino- and N-acylaminopyridiniums produce nitrenium ions on thermolysis. Thus, heating 1-(N-phthalimido)-2,4,6-triphenylpyridinium tetrafluoroborate in mesitylene/(CF₃)₂CHOH at 180�C gave N-mesitylphthalimide, while thermolysis or photolysis in anisole gave phthalimidoanisoles in 71.6% yield.^4b Photolysis of 1-(N-phthalimido)-2,4,6-trimethylpyridinium tetrafluoroborate in anisole containing (CF₃)₂CHOH and CF₃CO₂H gave, following hydrazinolysis, the anisidines (94.5%) (o-: m-: p- :: 66.4: 8.3: 20.3), clearly suggesting that an electrophilic reactive intermediate (the diacylnitrenium ion) was generated.

    The 1-(1,2,4-triazol-4-yl)-2,4,6-trimethylpyridinium tetrafluoroborate (2) was readily prepared in 68% yield from 2,4,6-trimethylpyrilium tetrafluoroborate and 4-amino-1,2,4-triazole (3). Thermolysis of (2) under N₂ in degassed mesitylene (containing sufficient hexafluoroisopropanol to bring (2) into solution at room temperature) at 208�C for 7 h gave a mixture of at least 8 compounds, from which could be isolated 3-(1,2,4-triazol-4-yl)-2,4,6-trimethylpyridine (4) (3.9%), 4-(1,2,4-triazol-4-ylmethyl)-2,6-dimethylpyridine (5) (4.7%), 4-(2,4,6-trimethylbenzyl)-2,6-dimethylpyridine (6) (20.9%), and 1,2,4-triazole (7) (60.5%). No 2-(1,2,4-triazol-4-yl)mesitylene (8) was detected. An authentic sample of the latter was synthesized (44% yield) from N,N-dimethylformamide azine (9) and 2-aminomesitylene in the presence of p-toluenesulfonic acid.^6b

    The structure of (4) was established unambiguously by single crystal X-ray diffraction (Fig. 1).

Fig. 1. Thermal ellipsoid plot (50% probability) of 4 showing the atom numbering scheme.

    An alternate route to the 4-(1,2,4-triazoyl) cation was explored, namely the diazotization of 4-aminotriazole.⁹ Treatment of (3) with nitrosonium tetrafluoroborate in dry acetonitrile gave only (7), and no Ritter reaction-type products were detected. When the same reaction was carried out in the presence of mesitylene at room temperature, (7) was again isolated (77%) and nitrogen and nitrous oxide (the latter detected by infrared spectroscopy) were evolved. Remarkably, 1-(1,2,4-triazol-1-yl)-2,4,6-trimethylbenzene (10) was also isolated (1%), as were traces of 1-(1,2,4-triazol-4-yl)-2,4,6-trimethylbenzene (8) and the unsymmetrical mesitylene dimer (12). An authentic sample of (10) was synthesized from 2,4,6-trimethylphenylhydrazine hydrochloride and s-triazine using the general procedure of Grundmann and Ratz.¹¹ The same products, in approximately the same ratio, were obtained when (3) was treated with isoamyl nitrite in the presence of acetic acid and excess mesitylene, or when 1-amino-1,2,4-triazole (11) was treated as above with NO⁺BF₄^-. No (10) was detected in the thermolysis of (2) in mesitylene.

    The formation of the 1-yl isomer (10) in reasonable amounts in the photolysis of (2) is noteworthy, compared with the formation of the 4-yl products (4) and (5) in the thermolysis (no (10) detected). This suggested the possibility that, in the photolysis, a concerted heterolysis of the pyridinium salt was assisted by the nucleophile mesitylene attacking the triazole at N-1. To test this possibility, an isomer of (2), 1-(1,2,4-triazol-1-yl)-2,4,6-trimethylpyridinium tetrafluoroborate (17) was synthesized from (11) and 2,4,6-trimethylpyrilium tetrafluoroborate. Its photolysis in mesitylene/actonitrile gave (10) (45 %) and traces of (8) and (12) (detected by GC/MS), in addition to (7) (19 %), thus negating this possibility.

    The photolysis was carried out using a variety of solvent combinations (acetonitrile, hexafluoroisopropanol, mesitylene) and reaction conditions (degassed, in the presence of air, at various temperatures) and using either 254, 300, or 350 nm lamp. The highest yield of (10) (48%) was obtained from (2), using mesitylene in anhydrous degassed acetonitrile at 0 - 5 �C and the 300 nm lamp.

    When (2) was photolyzed in pure acetonitrile, the major products were (7) (6%), 2,4,6-collidine (12%), and (13) (20%), with traces of (4) (GC/MS) (Table 1). To determine whether the formation of (13) involved an intra- or intermolecular process, the triphenylpyridinium salt was photolyzed in acetonitrile containing 2,4,6-collidine in excess and gave mainly (13). The corresponding photolysis of (2) in acetonitrile/2,6-lutidine gave a mixture of two isomers of (18), in addition to some (7) and (13).Thus, the formation of (13) is clearly an intermolecular process in the case of the above photolysis of the triphenylpyridinium, as is the formation of (18) from (2), suggesting that the formation of (13) from the trimethylpyridinium (2) in mesitylene is also an intermolecular process. The 4-yl isomers (4) and (8) are not photoisomerized to the corresponding 1-yl compounds (13) and (10), though (4) and (8) are slightly decomposed after irradiation with 300 nm light in degassed acetonitrile for 20h.

Table 1. % Yields of Products Formed from the Photolysis of (2) in Various Solvents.^a

    We also carried out experiments to see whether or not a 1,2,4-triazolyl free radical would react with mesitylene or with 2,4,6-collidine to form (8), (10), (4), or (16). Photolysis of (21) in mesitylene using 300 nm light¹⁷ proved to be a clean source of the desired radical, producing the products expected for such a radical: 1,2,4-triazole (7), dibenzoylimide (22), and the symmetrical 3,5-dimethylbenzyl radical coupling product (23). No (10) was detected.¹⁸

Fig. 2. Selected geometrical parameters of the CASSCF(10,8)/6-31G(d) optimized structures for the most relevant low-lying electronic states of the 1,2,4-triazolyl cation. Distances are given in angstroms and angles in degrees.

Fig. 3. Simple qualitative pictorial descriptions of the four highest occupied MOs of the most relevant low-lying electronic states of the 1,2,4-triazolyl cation.

Table 2. Total Energies (E, Hartrees) Calculated at Different Levels of Theory and Zero-Point Vibrational Energies (ZPVE, kcal/mol) for Low-Lying Electronic States of 4-(1,2,4-Triazolyl) Cation

Thermolyses.

    The most obvious difference between the thermolyses and photolyses of the N-triazolopyridiniums is that the triazole moiety does not lead to a triazolomesitylene in the former, but does in the latter. The absence of (8) in the thermolysis mandates the absence of cation (1) as a free entity under these conditions. The most likely route to (5) would go via the anhydro-base (24): a 1,5-sigmatropic shift (thermally allowed) would lead to (5). A similar 1,5-shift (also thermally allowed since the positive charge is not involved in the reorganization of the electrons) would lead to (4). We suggest that (6) may be formed from (24) by an attack of mesitylene at the methene carbon concerted with the departure of the triazole anion (Scheme 1).

    To explain the formation of (7) as the main product requires a competing process, since triazole could not be formed in a concerted process. One possibilty is that (2) undergoes N-N bond homolysis giving the triazolyl free radical (see later), which abstracts a benzylic hydrogen to give the observed (7), and a benzyl-type radical which will dimerize. Two facts argue against this proposal: there seems to be no precedent of such a homolysis taking place with the pyridiniums,¹ and secondly, no products resulting from benzylic radical coupling were observed in this case. Had free (1c) been formed by N-N heterolysis, one would have expected at least some electrophilic attack on mesitylene to give a 1-triazolyl substitution product (as was observed in the diazonium ion decompositions). Since no such product was formed in the thermolyses, we suggest that the triazolyl cation forms a tight pair with collidine, which then undergoes a SET, leading to the collidine radical cation (25) and the triazolyl radical (26) (again within the tight pair). Abstraction of a hydrogen atom by triazolyl radical (26) from the 4-methyl group of (25) could lead to (7) and the 4-picolyl cation (27). Such benzyl-type cations are known to give diarylmethanes,^12a,20 while the corresponding benzyl free radicals dimerize to give 1,2-diarylethanes.²⁰ Thus, a Friedel-Crafts alkylation of mesitylene by (27) might account for the formation of (6). Hydride ion abstraction from (2) leading to (24), and formation of (6) from (24) will both lead to (7), as could other hydrogen-abstraction processes. That (4) and (5) (both 4-triazolyl compounds) but no 1-triazolyl compounds are formed in the thermolysis of the pyridiniums argues again in favor of a concerted mechanism leading to these products. On the other hand, (6) and (7) could well result from the heterolytic mechanism.

Photolyses.

    In contrast, the results from the photochemical reactions indicate that the products are probably formed by intermolecular stepwise processes. Some nitrenium ions have recently been detected using laser flash photolysis.^13a,21 We have carried out very preliminary experiments on the laser flash photolysis (355nm, 6mJ) (30 ps pulse) of (2) and the corresponding 2,4,6-triphenylpyridinium salt. Both salts visibly fluoresce and show transients upon direct excitation. The absorbance maximum of the transient from (2) is at 485 nm (Fig. 4), and those from the triphenyl compound are at 450 and 630 nm (Fig. 5).

    The transient from (2) decays on the timescale of a few nanoseconds, which decay is somewhat accelerated by the addition of mesitylene. The position and shape of the spectral band, shown in Fig. 4, is not altered by the presence of mesitylene. The spectrum arising from the triphenyl compound changes between 50 and 500 ps following the laser pulse, with the 450 nm peak decaying and producing a new absorption between 490 and 600 nm.

    Some credibility is lent to this hypothesis by the finding by Katritzky and coworkers that 1,2-diphenylpyridinium perchlorates undergo intramolecular ring-closure on irradiation with 300 nm light,²³ and 2-phenyl-1-(1,2,3-triazol-4-yl)pyridinium salts undergo similar intramolecular cyclizations.²³ The above results, along with the reported photochemical properties of pyridinium compounds in the presence of electron donors,²⁴ led us to examine the possibility of a photoinduced electron transfer (PET) from mesitylene to salt (2). The intensity of fluorescence of (2) as a function of mesitylene concentration was measured and no change was observed in the shape, but a change was observed in the intensity of the fluorescence leading to the Stern-Volmer relationship²⁴ displayed in Figs. 6a,b. The quenching rate constant, k_q (= K_qt^-1), in acetonitrile, obtained from the slope (K_q) of the Stern-Volmer plot and the fluorescence lifetime t (ca. 2 ns, from the LFP studies), is 1.4 x 10¹⁰ M^-1s^-1. A similar value (1.3 x 10¹⁰ M^-1s^-1) has been reported for electron transfer from mesitylene to 10-methylacridinium ion.²⁵ Such an electron transfer would produce the mesitylene radical cation (29) and the pyridinyl radical (30), which could fragment to the triazolyl radical (26) and then lead to (10) (Scheme 3), as suggested for the formation of (4) in Scheme 1 above.

    Pyridinyl radicals, which are accessible by electrochemical reduction, photolysis,^26a or by chemical means,^26b are isolable in some cases,^26c but usually dimerize or fragment. The latter is the predominant reaction observed in cases in which a heteroatom is bound to the pyridinyl nitrogen,^26c such as is (30). To test if (30) could be a possible intermediate in our reactions, (2) was reduced with Na/Hg under nitrogen in acetonitrile solution.²⁷ 2,4,6-Collidine and triazole (7) were the major products, together with lesser amounts of collidine dimers. No detectable amounts of the dimer of radical (30) were detectable (NMR, GC/MS). This fragmentation of (30), together with the LFP spectroscopic data, strongly suggest that (10) is not formed via the intermediacy of (1).

    On the other hand, the formation of (1c) as a tight pair with either 2,4,6-collidine or 2,4,6-triphenylpyridine may be required to explain to account for the formation of (13) (with traces of (4)) from (2) in the absence of mesitylene (Scheme 4). In the presence of added 2,6-dimethylpyridine this tight pair may be solvent-separated leading to the formation of (18) and (7) ((13) could still arise from the tight pair). A similar separation of the pair from the triphenylpyridinium in the presence of external collidine would explain the products similarly (here, (13) would necessarily arise from the solvent separated pair).

    That no product resulting from hydrogen-abstraction from the pyridine side-chains was observed may be due to the fact that, if this abstraction did occur, it would produce a picolyl cation in which the positive charge would be delocalized over the nitrogen atom (i.e. a nitrenium ion) -- see argument above in the discussion of the thermal formation of (6).

    The thermolysis of the diazonium salts from (3) and from (11) produces 77% of the parent triazole (7), a result consistent with the formation of the diradical cation (1c) (capable of hydrogen abstractions to form triazolium tetrafluoroborate, which provides (7) upon isolation). Also formed was N₂O, which is consistent with either the formation of a hydroxyazo intermediate²⁸ or a labile N-nitrosoamine,²⁹ and the mesitylene dimer (12). The triazolomesitylenes (10) and (8) are formed from both (3) or (11) (in very low yield) in about the same ratio. This suggests the formation of a common intermediate. If mesitylene were reacting directly with the two diazonium salts (3) and (11) via an S_N2' process, different ratios of (10) to (8) would be expected to be formed from the two aminotriazoles, so this can be excluded. That the reactive intermediate from the thermolysis is not a free radical³⁰ was established by the photolysis of (21) (which gives rise to the radical) in the presence of mesitylene, but no (10) was formed in this way. The 1- and 4-triazolyl cations could produce (10) via electrophilic attack of mesitylene or via an SET process similar to that suggested for the thermolysis of (2). The hydrogen-abstraction product (7) must then arise from the triplet state (1c) of the diradical cation, which is predicted to be the ground state of (1).

    Unlike other nitrenium ions (singlets) generated photolytically from the corresponding pyridiniums,^2e,14a the triplet diradical cation (1c) does not react in the same way with electron rich donors, but gives mainly the hydrogen abstraction product from suitable donors, as is typical of triplet nitrenium ions. Note in Table 1 the increase in yield of triazole (7) from 6% to 42% as the methanol-d₄: acetonitrile ratio increases. In conclusion, therefore, it appears that the thermolysis of the pyridinium salt (2) does not generate free (1) and the products (with the possible exception of triazole itself) result mostly from concerted processes. Thermolysis of the N-diazonium salts produces (1), while photolysis of the pyridiniums in the presence of mesitylene gives products whose formation is best accounted by single electron transfer from the substrate to the pyridinum, whereas photolysis in methanol and acetonitrile is best explained by involving the intermediacy of (1). Finally, the triazolium's behavior (when it is formed) is well-accounted for by its predicted diradical cation structure (1c).

Acknowledgments.

(2) (a) Abramovitch, R. A.; Cooper, M.; Iyer, R.; Jeyaraman, R. Rodrigues, J. A. R. J. Org. Chem. 1982, 47, 4819. (b) Abramovitch, R. A.; Chinnasamy, P.; Evertz, K. Huttner, G. J. Chem. Soc., Chem. Commun. 1989,3. (c) deSousa, J. D. F.; Rodrigues, J. A. R.; Abramovitch, R. A. J. Am. Chem. Soc. 1994, 116, 9745. (d) Abramovitch, R. A.; Ye, X-C.; Pennington, W. T.; Schimek, G.; Bogdal, D. J. Org. Chem. 2000, 65, 343. (e) Takeuchi, H. J. Chem, Soc., Chem. Commun. 1987, 961.

(3) Novak, M.; Xu, L.L.; Wolf, R. A. J. Am. Chem. Soc. 1998, 120, 1643. Novak, M.; Kennedy, S. S. 1995, 117, 574. Humphreys, W. G.; Kadlubar, F. F.; Guengerich, F. P. Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 8278. Colvin, M. E.; Nilsen, I. B. M.; LeBui,L.; Hatch, F. T. Chem. Biol. Interact. 1992, 81, 19, and references cited therein.

(4) (a) Abramovitch, R. A.; Evertz, G.; Gibson, H. H., Jr.; Weems, H. G., Jr. J.Chem. Soc., Chem. Commun 1988, 325. (b) Abramovitch, R. A.; Beckert, J. M.; Chinnasamy, P.; He, X-H.; Pennington, W. T.; Sanjivamurthy, A. R. V. Heterocycles 1989, 28, 623

(5) Rutchenko, V.; Ignatov, S.; Kostyanovsky, R. J. Chem. Soc., Chem. Commun. 1990, 261.

(6) (a) Abramovitch, R. A.; Beckert, J. M.; Pennington, W. T. J. Chem. Soc., Perkin Trans. 1 1991, 761. (b) Gilchrist, T. L.; Rees, C. W.; Thomas, C. ibid. 1975, 12.

(7) Fabian, J.; Melhon, A.; Tynkylkov, N. Theochem. 1987, 151, 355.

(8) Gompper, R.; Junius, M. Terahedron Lett. 1980, 21, 2883. Gompper, R.; Guggenberger, R.; Zantgraf, R. Angew. Chem. Int. Ed. Engl. 1985, 24, 984.

(9) Abramovitch, R. A.; Gibson, H. H.; Nguyen, T.; Olivella, S.: Sole, A. Tetrahedron Lett. 1994, 35, 2321.

(10) Li, Y.; Abramovitch, R. A.; Houk, K. J. Org. Chem. 1989, 54, 2911. Ford, G. P.; Herman, P. S. J. Chem. Soc., Perkin. Trans. 2 1991, 616. Smith, D. A.; Bitar, J. J. Org. Chem. 1993, 58, 6.

(11) Grundmann, C.; Ratz, R. J. Org. Chem. 1956, 21, 1037.

(12) (a) Wen, L.-S.; Kovacic, P. Tetrahedron 1978, 34, 2723. Kovacic, P.; Wu, C. J. Org. Chem. 1961, 26, 749. (b) See, however, Takeuchi, H. J. Chem. Soc., Chem. Commun. 1987, 961, and discussion in footnote to ref. 4a above.

(13) (a) Moran, R. J.; Cramer, C;, Falvey, D. E. J. Org. Chem. 1997, 62, 2742. (b) The known reactivity of acetonitrile towards cations, radicals and radical cations may explain, in part, the low material balance in some of our reactions, although we did not detect any Ritter reaction products or stable solvent-derived products. See Lisjer, H. J. P.; Arnold, D. R. J. Phys. Chem. A 1998, 102, 5592.

(14) Fleming, I.; Dunogues, J.; Smithers, Organic Reactions 1989, 37?57.

(15) Ohga, K.; Yoon, U. C.; Mariano, P. S. J. Org. Chem. 1984, 49, 213.

(16) The low yield of 19 may also be owing to the reported desilylation of the allylsilane-derived radical cation by the weakly nucleophilic acetonitrile. Ohga, K.; Mariano, P. S. J. Amer. Chem. Soc. 1982, 104, 617.

(17) See the similar generation of the 1,2,3-triazolyl radical. Kalambokis, E. A.; Maroulis, A. J.; Alexandrou, N. E. J. Heterocycl. Chem. 1993, 30, 1301.

(18) t-Butyl 1-(1,2,4-triazolepercarboxylate), synthesized from t-butylhydroperoxide and 1,1'-carbonylbis-(1,2,4-triazole), underwent nucleophilic attack at the carbonyl group, as does t-butyl 1-imidazolepercarboxylate. Bourgeois, M. J.; Filliatre, C.; Lalandem R.; Maillard, B.; Villenave, J. J. Tetrahedron Lett. 1978, 36, 3355.

(20) Kozhevnikov, I. V.; Kim, V. I.; Talzi, E. P.; Sidelnikov, V. N. J. Chem. Soc., Chem. Commun. 1985, 1392. Eberson, L.; Nyberg, K. Accts. Chem. Res. 1973, 6, 106. Abramovitch, R. A.; Roy, J.; Uma, V. Can. J. Chem. 1965, 43, 3407.

(21) (a) Moran, R. J.; Falvey, D. E. J. Am. Chem. Soc. 1996, 118, 8965. (b) Chiapprino, D.; Falvey, D. E. J. Phys. Org. Chem. 1997, 10, 917. (c) Srivastava, S.; Kercher, M.; Falvey, D. E. J. Org. Chem. 1999, 64, 5833.

(22) Wagner, B. D.; Ruel, G.; Lusztyk, J. J. Amer. Chem. Soc. 1996, 118, 13.

(23) (a) Katritzky, A. R.; Zakaria, Z.; Lunt, E. J. Chem. Soc., Perkin Trans.1 1980, 1879. (b) Katritzky, A. R.; Agha, B.; deVille, G. Z. Org. Mag. Resonance 1983, 21, 649.

(24) (a) Mariano, P. Accts. Chem. Res. 1983, 16, 130. (b) See also: Bockman, T. M.; Lee, K.H.; Kochi, J.; J. Chem. Soc., Perkin Trans. 2 1992, 1581.

(25) Fujita, M.; Ishida, A.; Takamuku, S.; Fukuzumi, S. J. Amer. Chem. Soc. 1996, 118, 8566.

(26) (a) Kosower, E.M.; Schwinger, I. J. Amer. Chem. Soc. 1964, 86, 4493. (b) Kosower, E. M. �Molecular Biochemistry,� McGraw Hill, N.Y. 1962. (c) Kosower, E. M.. "Free Radicals in Biology," (Pryor, W. A., Ed.), Vol. II, Ch. 1, Academic Press, New York, 1976.

(27) Kosower, E. M.; Ikegami, Y. J. Amer. Chem. Soc. 1967, 89, 461.

(28) (a) Kreher, R.; Bergmann, V. Z. Naturforsch., Anorg. Chem., Org. Chem. 1992, 37b, 881; (b) DeRosa, M.; Haberfield, P. J. Org. Chem. 1981, 46, 2639, and references cited therein.

(29) Kuzmenko, V. V.; Pozharskii, A. F. Adv. Heterocycl. Chem. 1992, 53, 154.

(30) March, J. �Advanced Organic Chemistry, Reactions, Mechanisms, and Structure, 4th Ed.,� Ch. 14, Sec. 4-18, John Wiley & Sons, New York, 1992.