This follows on from my answer to your previous question: Factors on which Coefficient of restitution depend. The coefficient of restitution of a collision depends on the available degrees of freedom for energy to be lost.
If you take your example of the collision of atomic particles, let's say two electrons, then there isn't anywhere for the initial kinetic energy to go. You can't make an electron vibrate, or rotate, or plastically deform it, so all the kinetic energy you put in has to come back out as kinetic energy and the coefficient of restitution is unity.
Though actually electron collisions are not perfectly elastic. Even at low energy the collision will radiate some photons and energy will be lost as a result. That means the coefficient of restitution will be (slightly) less than one. Raise the initial energy to the 209 GeV used in the LEP collider and lots of energy will be lost in creating other particles so the coefficient of restitution may be much less then one.
The reasons why collisions of macroscopic objects have coefficients of restitution less than unity were discussed in my answer to your previous question.
I don't know what you mean by perfectly elastic collision occur during shooting or super elastic collision occur during explosion. If by the latter you mean why does the shrapnel end up with more kinetic energy than it started with, that's because chemical energy of the explosive was converted to kinetic energy. However I think calling this a collision is stretching things a bit.