I used also iron, nickel, zinc, copper. However, cobalt gives the easiest and straight forward yield of nitrides (XPS and SAED verified). It is so easy to get cobalt nitride that I don't even have to use high purity nitrogen as I need with $\ce{Fe},$ $\ce{Ni},$ $\ce{Zn},$ $\ce{Cu}.$
Is there any reason for that? Is it because of back-bonding?
According to the recent first principles ab initio DFT calculations done with VASP of structural, electronic and mechanical properties of 3d transition metals mononitrides $\ce{MN}$ carried out by Liu et al. [1], there is a local maximum of cohesion energy $E_\mathrm{coh}$ at cobalt for all possible/discovered structure types:
Ultra-soft Vanderbilt pseudo-potentials (US-PP) … as supplied by Kresse and Hafner …, with local density approximation (LDA) … and PW91 general gradient approximation (GGA) … were both used. It is known that LDA underestimates the lattice constant and overestimates the elastic constants and cohesive energy, while PW91 GGA does the opposite for the lattice constant and elastic constants by a smaller amount and yields more accurate values of cohesive energy.
Figure 1. Cohesive energy per atom of the nitrides $(\ce{MN})$ versus their corresponding 3d transition metals $(\ce{M}).$ The three structures are zincblende (zb, black square), rocksalt (rs, red circle) and cesium chloride (cc, blue triangle). Experimental values for the rocksalt structure (green ×) … are also presented.
Since we are discussing synthesis from the elements, it's especially handy to use cohesion energy since for solids it signifies the energy required to disassemble the structure into an array of neutral free atoms. As a rule of thumb, the higher the atomic cohesion energy is, the more stable the solid is. Numerical values of $E_\mathrm{coh}^\mathrm{GGA}$ for cobalt nitrides are always exceeding the ones for iron, nickel, copper and zinc, and also $\ce{CoN}$ demonstrates mechanical stability across all structure types:
There is still a lot of controversy between the studies of preferred phase and magnetism [2]; I'm not sure though it is currently possible to interpret/rationalize or let alone predict this behavior by other, simplified means.
I think this property has to do with the ease with which cobalt makes complexes with $\ce{NH3}$ and amines. The variety and the stability of the complexes containing $\ce{Co^3+}$, surrounded by $\ce{NH3},$ ethylenediamine, $\ce{H2O},$ and/or $\ce{Cl},$ is unparalleled. Iron $\ce{Fe}$ does not make complexes with $\ce{NH3}$ under usual conditions. Nickel, copper and zinc do make one or two complexes with $\ce{NH3},$ but they are easily destroyed in acidic conditions. The variety and the stability of the cobalt complexes is unequaled.
The stability of the cobalt complexes is related to its electronic structure. The cobalt atom is argon + 9 electrons. $\ce{Co^3+}$ is argon + 6 electrons. 12 electrons are missing in $\ce{Co^3+}$ and needed to build up the electronic configuration of the next noble gas, krypton. If the $\ce{Co^3+}$ ion is surrounded by 6 ligands like $\ce{NH3},$ $\ce{Cl-}$ and $\ce{H2O},$ it "achieves its goal" of looking like a noble gas.
Cobalt may react with nitrogen to produce $\ce{CoN}$ made of $\ce{Co^3+}$ and $\ce{N^3-}$. And maybe this compound $\ce{CoN}$ is surrounded by six $\ce{N2}$ in the same way as $\ce{Co^3+}$ is surrounded by six ligands in the $\ce{Co^3+}$ complexes. This should be checked by X-ray analysis.
Metal nitrides can be produced from the amide ($\ce{NH2−}$) ion as the $\ce{N3−}$ source (Metal nitrido complex).
Also, heating water vapor to high temperature:
$$\ce{H2O ->[\Delta] .H + .OH}$$
$$\ce{.H + .H -> H2}$$
In the presence of atomic nitrogen from heated nitrogen:
$$\ce{.N + H2 -> .NH2}$$
Cobalt is an active transition metal which can assume a variety of oxidation states which facilitates the release of electrons:
$$\ce{Co(K) -> Co(K + 1) + e-}$$
Finally,
$$\ce{.NH2 + e- -> NH2-}$$
creating the favorable amide ion which is a reported source for $\ce{N3−}$.
