Far-field divergence

Diffraction ties spot size to spread: the smaller the emitting aperture, the faster the beam diverges. An edge-emitting laser's mode is confined to roughly a quarter-micron vertically and a few microns laterally, so its far field is broad, elliptical, and oriented counter to intuition. It is widest in the axis where the near field is narrowest.

Convention quotes two FWHM angles: $\theta_\perp$ (fast axis, perpendicular to the epitaxial layers) and $\theta_\parallel$ (slow axis). Representative edge-emitter values: $\theta_\perp \approx 25{-}40°$, $\theta_\parallel \approx 6{-}12°$ (a 3–5:1 ellipse). VCSELs, emitting from a circularly symmetric aperture several microns across, produce nearly round beams of ~10–25° and are the easy case for coupling.

Three consequences drive optical design:

Collection NA. Capturing a 35° FWHM fast axis without truncation demands lens NA ≳ 0.5 once full-power widths are honored. Truncation costs power and stamps diffraction structure onto the beam. This is why laser collimators are high-NA aspheres, not ordinary achromats.

Ellipticity management. Downstream systems mostly want round beams; anamorphic prism pairs, cylindrical lenses, or beam-shaping micro-optics rescale one axis.

Astigmatism. In gain-guided and some weakly index-guided structures, the apparent source positions for the two axes differ longitudinally by microns to tens of microns; a single spherical lens then cannot focus both axes simultaneously. Astigmatism must be measured and compensated (weak cylinder) for tight-focus applications.

Far fields are measured with goniometric scans or calibrated camera systems at large distance; report FWHM versus $1/e^2$ conventions explicitly, since the same beam yields materially different numbers ($1/e^2$ ≈ 1.7× FWHM for a Gaussian). Divergence also feeds directly into fiber-coupling estimates through mode-field matching.