What the two models include. Simple is the standard treatment — Fresnel coupling into the crystal, dispersion of the electro-optic coefficient through the Faust–Henry relation, the phonon's own absorption, velocity mismatch integrated over the thickness, and a fixed probe gate. All effects adds absorption raised to measured levels (multiphonon and residual, not just the phonon tail), probe group-velocity dispersion propagated through the crystal, and Fabry–Pérot round trips of the terahertz pulse.

In practice the etalon dominates. Absorption and dispersion move the answer by a few per cent; the internal reflection sends about 29 % of the terahertz intensity back on every bounce and puts ripple across the whole spectrum. Bonding the active crystal to a thick inactive handle pushes that echo outside the scan window, which is why detector crystals are usually supplied as bonded pairs.

Not included. Two-photon absorption of the probe, which matters for GaP at 800 nm and depends on your intensity; the probe's own etalon and its cross terms with the terahertz echo; over-rotation, which sets in well before the field strengths an intense source delivers. Treat the curve as the linear, small-signal response.

Material constants are room-temperature literature values for ⟨110⟩ crystals at normal incidence. Absorption above the phonon is extrapolated. Swiss Terahertz supplies GaP and ZnTe detector crystals — ask us for the grade and bonding that suits your setup.