Optical anisotropy and crystal symmetry
In an isotropic medium such as glass, the speed of light is the same regardless of propagation direction or polarisation, and a single refractive index n describes the medium completely. Many crystals, however, are optically anisotropic: atoms are arranged differently along different crystallographic axes, so the restoring forces on bound electrons — and hence the refractive index — depend on both the direction of propagation and the polarisation of light. Cubic crystals (diamond, halite) stay isotropic with a single index; trigonal, tetragonal and hexagonal crystals are uniaxial with two principal indices no and ne; orthorhombic, monoclinic and triclinic crystals are biaxial with three. Erasmus Bartholin first described double refraction in calcite in 1669; Christiaan Huygens gave a wave-mechanical explanation in 1690, nearly a century before polarisation was formalised.
Ordinary and extraordinary rays
When unpolarised light enters a birefringent crystal at an oblique angle, it splits into two refracted beams. The ordinary ray (o-ray) obeys Snell's law exactly, with spherical wavefronts and refractive index no. The extraordinary ray (e-ray) does not obey Snell's law in the usual form — its index ne(θ) varies continuously with the angle θ between the ray and the optical axis, and its wavefronts are ellipsoidal. The two rays are polarised perpendicular to each other and exit the crystal at slightly different positions, producing calcite's famous doubled image.
1/n_e(θ)² = cos²(θ)/n_o² + sin²(θ)/n_e² Calcite: n_o = 1.6584, n_e = 1.4864 → Δn = -0.172 (negative uniaxial) Quartz: n_o = 1.5443, n_e = 1.5534 → Δn = +0.009 Rutile: n_o = 2.616, n_e = 2.903 → Δn = +0.287
The optical axis and retardation colours
The optical axis is the one crystallographic direction along which both polarisation states travel at the same speed, so birefringence vanishes exactly along it. When a birefringent slab of thickness d sits between crossed polarisers, the accumulated path difference Γ = (no − ne)·d = Δn·d selects which wavelengths transmit and which are extinguished, producing the vivid Michel-Lévy interference colours used by petrographers to identify minerals. Colours progress from first-order grey and yellow through second-order greens and pinks as retardation increases, eventually washing out to pastel higher orders.
Wave plates: converting between polarisation states
A wave plate is a birefringent slab cut so its optical axis lies in the plate surface. A quarter-wave plate introduces a Γ = λ/4 (90°) retardation, converting linear polarisation at 45° to the axis into circular polarisation — used in optical isolators and LCD backlights. A half-wave plate introduces a 180° shift and rotates the plane of linear polarisation by twice the plate's orientation angle, a tool for adjusting laser polarisation without reflective losses. Two birefringent prisms cemented with perpendicular axes form a Wollaston prism, the basis of differential interference contrast microscopy.
From LCD pixels to gemstone identification
Every pixel of a twisted-nematic LCD is a voltage-controlled birefringent cell between crossed polarisers. Polarising microscopy uses interference colours and conoscopic figures to identify rock-forming minerals and map crystal orientation. Nonlinear optical crystals (KDP, KTP, BBO) exploit birefringence for phase-matched second-harmonic generation, converting infrared laser light to visible or UV. Amorphous glass and plastics become birefringent under mechanical stress — photoelasticity — letting engineers visualise stress concentrations directly, and a doubled image seen through a gemstone's back facets is a standard diagnostic in gemology.
Frequently asked questions
Why does a calcite crystal produce two images?
Calcite is strongly birefringent: it splits any entering light beam into an ordinary ray and an extraordinary ray that obey different refraction laws and thus exit the crystal at slightly different lateral positions. Because the two exit points are spatially separated, two distinct images of any object behind the crystal are formed. Rotating the calcite causes one image to orbit the other, tracing the rotation of the extraordinary ray direction around the optical axis.
What is the optical axis of a crystal?
The optical axis is a specific crystallographic direction along which both allowed polarisation states of light travel at exactly the same speed, so birefringence is zero for propagation along it. Uniaxial crystals have one such axis; biaxial crystals have two. Light travelling along the optical axis sees no double refraction, while light travelling perpendicular to it experiences maximum birefringence.
How is birefringence used in LCD screens?
Each pixel of a twisted nematic LCD is a voltage-controlled birefringent cell. Without voltage, liquid crystal molecules rotate the backlight polarisation by 90 degrees, allowing it to pass through the front polariser for a bright pixel. Applying a voltage unwinds the helix, the polarisation is no longer rotated, and the polariser blocks the light for a dark pixel. Intermediate voltages produce grey levels.
Try it live
Everything above runs in your browser — open Birefringence — Double Refraction in Crystals and watch ordinary and extraordinary rays split inside a crystal, then rotate it and adjust thickness to see the interference colours change between crossed polarisers. Nothing is installed, nothing is uploaded.
▶ Open Birefringence simulation