Gain compression factor (ε)

The differential gain a laser mode experiences is not constant as photon density rises: carrier heating, spectral hole burning within the gain spectrum, and carrier transport effects all reduce the available gain on picosecond timescales. Rather than model each microscopic mechanism, laser dynamics absorbs them into one phenomenological parameter, the gain compression factor $\varepsilon$:

with $S$ the photon density. Typical values for quantum-well lasers sit around $\varepsilon \approx 1\text{--}5 \times 10^{-17}\ \mathrm{cm^3}$ \u2014 small enough that $\varepsilon S \ll 1$ in normal operation, large enough to dominate the laser's damping.

The practical consequence lives in the modulation response. Gain compression contributes the photon-density-proportional term in the damping rate $\gamma$ of the relaxation oscillations: pushing a laser harder raises the resonance frequency $f_r$ but also raises damping, and the K-factor \u2014 the slope of $\gamma$ versus $f_r^2$ \u2014 is, to leading order, a direct measurement of $\varepsilon$ (plus the photon lifetime). This is why $\varepsilon$ appears whenever someone asks how fast a directly modulated laser can ultimately go: the maximum bandwidth scales inversely with K, hence with gain compression.

Above the small-signal picture, $\varepsilon$ also flattens the large-signal eye (compressing overshoot), shifts the clamped carrier density slightly upward with power, and couples into chirp through the accompanying index change.

$\varepsilon$ is extracted rather than looked up: fit damping versus resonance frequency from small-signal modulation or RIN spectra at several bias points. Quoted values vary between devices and between fitting conventions (some authors compress $g$, others compress $v_g a$) \u2014 when comparing papers, check which convention the rate equations used.