Two parallel partially reflecting mirrors (intensity reflectivity R each, no absorption) trap light. Each round trip adds a phase δ; the transmitted beams are the geometric series of multiply reflected partial waves and interfere.
Equations used
δ = 4π n L cosθ / λT(δ) = (1−R)² / (1 + R² − 2R cosδ) = 1 / (1 + F sin²(δ/2)), F = 4R/(1−R)²FSR(λ) ≈ λ² / (2 n L cosθ)Finesse ℱ = π√R / (1−R) FWHM = FSR / ℱ (exact: FSR·(2/π)·asin(1/√F))resonance order m = 2 n L cosθ / λ (T = 1 when m is an integer)
Numerical checks (live)
The page sums the multiple-beam series t = (1−R)Σ Rk eikδ term by term (independent of the closed Airy formula) at 12 wavelengths and compares |t|² with T. It also scans the spectrum around a peak and compares the measured FWHM with the exact formula. Status: ...
Honest simplifications
Lossless mirrors with flat, wavelength-independent R and no mirror phase shift; real coatings absorb and disperse.
Collimated plane wave and perfectly parallel mirrors; no mirror roughness or parallelism error (these limit real finesse).
Cavity lengths here are in nanometres to show a few orders in the visible range; real etalons are micrometres to centimetres.
The FSR shown is the first-order wavelength formula, accurate when FSR << λ; the ray bounce diagram and standing wave are illustrations, not a field solution.
Refractive index n is held constant with wavelength (no dispersion).