Photonic crystal

A photonic crystal does for photons what a semiconductor lattice does for electrons: periodic modulation of the medium creates a band structure, and with sufficient index contrast and the right geometry, a photonic band gap (a frequency window in which no propagating mode exists in some or all directions).

By dimensionality: 1D photonic crystals are multilayer stacks. Every DBR mirror and fiber Bragg grating is one, with a gap (stopband) along the stacking axis. 2D crystals, the workhorse of the field, are triangular lattices of air holes etched through a suspended semiconductor slab (period ~400–450 nm for telecom in silicon or InP), giving an in-plane gap for one polarization while total internal reflection confines vertically. 3D crystals with complete gaps exist (woodpile, inverse opal) but remain fabrication showpieces rather than tools.

The gap becomes useful through its defects, exactly as in semiconductors. Remove a row of holes: a W1 waveguide, guiding light through the gap. Its engineered dispersion includes slow-light bands whose group indices reach 30–100, enhancing modulation and nonlinearity per unit length (at the price of proportionally enhanced loss and narrowband operation). Remove or shift a few holes: a nanocavity (the L3 and heterostructure designs), confining light to a mode volume near the theoretical minimum $(\lambda/n)^3$ while measured $Q$ factors reach $10^6$–$10^7$ in silicon.

That $Q/V$ combination is the technology's core asset, because light–matter interaction rates scale with it: Purcell-enhanced emitters and single-photon sources, thresholdless-regime nanolasers (high spontaneous-emission factor β), cavity QED with quantum dots, femtojoule-class optical switches, and ultra-dense sensors. The corresponding liability is sensitivity: nanometer disorder in hole size and position scatters slow-light modes and caps real-world $Q$. That keeps photonic-crystal devices concentrated where their extremes are indispensable rather than in general-purpose PDKs.