Dehydroazepines from phenyl azide photolysis

Sundberg and co-workers in 1974 reported results from flash photolysis studies of the reaction of phenyl azide with secondary amines [26]. Irradiation of the azide in hexane solution produced an intermediate absorbing strongly at 366 nm and having a lifetime of about 5 ms. This intermediate was found to react with amines to give, after tautomerization, the 3H-azepine that is obtained in preparative scale reactions. On the basis of these results, De Graff et al. [26] concluded that the detected intermediate is not a nitrene but is a closed-shell intermediate, either benzazirine or dehydroazepine. 

It now seems incontestable that irradiation of phenyl azide at room temperature gives dehydroazepine. Once formed, dehydroazepine can react with itself and/or phenyl azide to give tarry polymer or with nucleophiles to give substituted 3f/-azepines. The rate of reaction of dehydroazepines with amines depends dramatically on substitution the 5-acetyl substituted compound reacts 10,000 times faster than does the 5-methoxy substituted dehydroazepine. At low concentration of phenyl azide, dehydroazepine itself has a lifetime of approximately 5 ms and, presumably, isomerizes to phenyl nitrene. We will have more to say about this point later. The achievement of positive structural identification of the reactive intermediate formed in the room temperature photolysis of phenyl azide permits the detailed characterization of phenyl azide photochemistry. Further consideration of this analysis will be aided by examination of the results from time-resolved experiments for other aryl azides. 

It is clear now that irradiation of phenyl azide at room temperature gives dehydroazepine. At high concentration of azide, the dehydroazepine polymerizes rapidly in competition with its slow isomerization to triplet phenyl nitrene. The major product formed from photolysis of phenyl azide under conditions where its quantum yield for disappearance is claimed to be greater than unity is poly-1,2-azepine [48], not azobenzene. Of course, the polymer does not elute from an HPLC, and analysis of reaction mixtures by chromatography will show only two components. 

If the photoinitiated autocatalytic chain decomposition outlined in Scheme 12 operates at all, it should be specially important in the photolysis of (4-nitro)phenyl azide where formation of the dehydroazepine is a minor path and the triplet nitrene is formed rapidly. Indeed, Waddell and co-workers [110] report that the quantum yield for disappearance of this azide is 434 when a 0.1 M solution in acetonitrile is irradiated. However, attempts to reproduce this experiment reveal a quantum yield for disappearance of (4-nitro)phenyl azide in acetonitrile solution of 0.7 that is independent of starting azide concentration from 0.001 to 0.1 M [45]. 

---