Happy New Year, everyone! I am back.
There are many types of reversible covalent interactions but the one I am most curious about these days is the one between two nitrogens. I learned about it from Professor Jim Wuest (http://www.wuestgroup.com) while visiting Université de Montréal just before Christmas. Jim and his lab leverage the reversible coupling between nitroso compounds to make materials of controlled porosity. Being a small molecule guy, I am intrigued by the pair of reactions I saw in his nice Chemical Reviews account (http://pubs.acs.org/doi/abs/10.1021/cr500520s). Why is it that a 6-membered ring is not formed in the naphthalene case? There are a couple of lessons about aromaticity and strain here, among other things. This fascinating type of N=N bonding also takes me back to some of my earlier claims about the rarity of heterocycle construction using heteroatom-heteroatom bond formation. Here it is on display. However, these heterocycles are metastable.
amphoteros 
These benzofurazanes (after removing N-oxide with PPh3 reduction) are actually pretty good isosteres of a nitroaromatic compound (with lower toxicity liabilities, but slightly explosive); several of them advanced into clinical candidate molecules and their scale up routes featured in Org Process R&D journal. Typically, they would be made by bleach oxidation of o-nitroanilines, or by nitroazide thermolysis. (I think one additional reason may have been that until relatively recently benzofurazanes were underrepresented in the patent literature)
This is a great point and I agree. But the trick is indeed to remove that extra oxygen (using PPh3 as you are saying). Otherwise these 5-membered rings are metastable…
Pushing electrons in the first case looks very unusual (as well as the structure itself)
Indeed, the mechanism is rather convoluted…