There is no stimulated emission in an FEL.
The SASE (Self-attuned spontaneous emission) process works as follows:
- A bunch of electrons is accelerated.
- The electrons pass a first undulator, creating synchrotron-source type light.
- Light and electrons free-stream together, since both move at the speed of light, there is a back-reaction of the light on the electrons. This is the self-amplification phase: The (large-ish) electron bunch is cut up in several hundred "microbunches", which are compressed to length-scales on the order of microns (thus the name).
- The micro-bunches pass a second undulator. Here, they also create synchrotron light, but the duration of the pulses is the length of the microbunch divided by the speed of light
$$ \Delta t \approx \frac{\mu m}{c} = \frac{10^{-6} m}{10^8 \frac{m}{s}} \sim 10^{-14} s. $$
This is how a FEL can create femtosecond timescale light bursts.
- Since each microbunch emits its light coherently (note: this is a spatial coherence, not a phase coherence) the peak intensities are comparable to solid-state lasers, which is why these machines are called Free-Electron "lasers", even though there is no "lasing" in the conventional sense.
All of this is completely based on classical electrodynamics.
There is no stimulated emission into a highly-populated bosonic state, as in conventional lasers.
What I describe above is the unseeded FEL process. One can replace step 2 with a conventional laser providing the intense light needed to induce microbunching.