If auto-ignition temperature is the temperature when the vapor molecules/atoms just achieve boom velocity, instantaneously as their vibration energy increases with temperature, then we have a good correspondence of temperature of the vapor given its molecular/atomic mass, and kinetic energy that was previously used to define temperature. As in
PV=nRT
we have instead,
T=density∗f(vboom)
where T is temperature and f(vboom) is some function of vboom. Consider,
KE≡density∗f(vboom)
so,
KE=12∗density∗v2boom
this would be kinetic energy per unit volume, but,
vboom=3.4354∗density∗1particlecountpertype
so,
KE=3.435422∗density3∗1(particlecountpertype)2
And in the case of basic elements,
KE=density3∗5.901(atomicnumber)2
where the density (cubed) of the elements changes with T. A plot of T vs vboom of various density will give f(vboom) or the function f(x), from which we may then infer KE=g(T) where g(T) is a function of T. Temperature T here then, as intended is a measure of the vibrational kinetic energy within the element.
This seemingly is a definition of temperature without the explicit involvement of temperature particles, derived from the vibrational kinetic energy of the atoms in an elements. A sentinel point in this definition is Tboom, the temperature at which vboom is achieved.
This is not melting, but it would be interesting if Tboom occurs before the melting point.
Or is it? Melting? vboom changes density and could account for the energy released during phase change. There is a vboom from solid to liquid, a different vboom from liquid to gas and yet another value of vboom for the gas as it disintegrate. Different vboom values for different densities of the element/compound.
Is vboom responsible for phase change? Phase change would then be nuclear. Boiling water for tea would then be nuclear.