Light meeting a flat interface between two materials reflects a fraction set by the Fresnel formula R = ((n₁−n₂)/(n₁+n₂))². A single quarter-wave coating cancels that reflection at one design wavelength via thin-film interference, but drifts badly off that wavelength. A moth-eye nanostructure — subwavelength cones tapering from substrate index down to air — instead builds a continuous, graded effective index that light sees as a smooth ramp rather than a sharp step, suppressing reflection broadband across the whole visible spectrum.
R(λ) = |r|², via transfer-matrix method
n_eff(z) = √( f(z)·n_sub² + (1−f(z))·n_air² )
f(z) = fill fraction of the cone at height z
- Cone height — taller cones taper the index more gradually, extending the low-reflectance band to longer wavelengths (needs height comparable to λ/2n for full effect).
- Array period — must stay below λ/n so the structure behaves as an effective medium rather than a diffraction grating; the warning shows exactly when that assumption fails.
- Substrate — a bigger index step (e.g. silicon, n=3.5) starts from far higher bare reflectance, so nanostructuring buys a much bigger absolute gain than on glass.
This is the mechanism behind anti-glare phone glass, moth-eye corneal nanostructures that inspired it, and — closest to the article — nanotextured silicon solar cells, where cutting front-surface reflection directly raises the fraction of sunlight that reaches the absorbing layer.