I really can't warn strongly enough not to use sodium azide for ANYTHING that doesn't explicitly require its use in the reaction, it inhibits cytochrome oxidase in a similar manner to cyanides, and has roughly the same toxicity of KCN, that, and it is highly explosive, its used in car airbags, along with iron compounds, sillica etc. as the gas source in airbags.
The reason for the other impurities (impurities to those people scrounging NaN3 from car airbags anyway
) is to prevent one of the decomposition products of alkali metal azides, because the group one azides are pretty unstable and detonate from impact or flame into nitrogen, and a cloud of finely dispersed hot alkali metal.
On another note, I once considered controlled thermal decomposition of rubidium/caesium azide as a possible way to prepare the metals, maybe starting by preparing the azide from the alkali carbonate and hydrazoic acid (caution, a toxic volatile acid similar to hydrocyanic acid)
then thermal decomposition of continuous SMALL amounts of the azide, the nitrogen forming an inert atmosphere for collection of the alkali metal.
What chance does this have of succeeding? I think it would work for all alkali metals except lithium, due to its formation of the nitride directly from combination.