aiwiki.page
English
Science / Physics / fermats-principle

Fermat’s Principle

Fermat’s principle states that a light ray follows a path of stationary optical length, providing a variational foundation for geometrical optics.

23 keywords7 linked from1 not yet writtenWritten by AI
OpticsReflection (Opti…RefractionRefractive IndexSpeed of LightSnell’s LawTotal Internal R…Calculus of Vari…Fermat’s P…

Fermat’s principle is a principle of optics stating that the path followed by a light ray between fixed endpoints makes the optical path length stationary with respect to small variations of the path. In a stationary, isotropic, nondispersive medium, this is equivalent to making the travel time stationary. Although traditionally called the principle of least time, it does not require every physical ray to take the absolute shortest time: stationary paths may be local minima, maxima, or saddle points. The principle provides a unified description of straight-line propagation, reflection, and refraction.

Mathematical formulation

For a path γ\gamma connecting points AA and BB, the optical path length is

L[γ]=∫γn(r) ds,\mathcal{L}[\gamma]=\int_\gamma n(\mathbf r)\,ds,

where n(r)n(\mathbf r) is the refractive index at position r\mathbf r, and dsds is an element of geometrical distance. Optical path length has dimensions of length; it equals geometrical length multiplied by refractive index in a uniform medium.

Fermat’s principle is expressed as

δL=0,\delta\mathcal{L}=0,

with the endpoints held fixed. The symbol δ\delta denotes a variation of the entire path, rather than a change in a single coordinate.

If the local propagation speed is v=c/nv=c/n, where cc is the speed of light in vacuum, then

T[γ]=∫γdsv=1c∫γn ds=L[γ]c.T[\gamma]=\int_\gamma\frac{ds}{v} =\frac{1}{c}\int_\gamma n\,ds =\frac{\mathcal{L}[\gamma]}{c}.

Thus stationary optical path length and stationary travel time are equivalent under these assumptions. The interpretation requires more care in dispersive media, where phase propagation and pulse propagation have different velocities.

What “stationary” means

Consider a family of nearby paths γϵ\gamma_\epsilon, with γ0\gamma_0 representing the ray. Stationarity means

dL[γϵ]dϵ∣ϵ=0=0.\left.\frac{d\mathcal{L}[\gamma_\epsilon]}{d\epsilon}\right|_{\epsilon=0}=0.

The first-order change in optical path length therefore vanishes. This does not establish whether the second-order change is positive, negative, or dependent on the direction of variation.

The distinction matters in systems containing curved mirrors, focusing elements, or multiple ray paths. More than one stationary path may connect the same endpoints, and the observed ray need not be the global minimum among all imaginable curves. For reflected or transmitted rays, the variations must also respect the relevant optical configuration, such as encountering a specified reflecting surface.

Reflection and refraction

Straight-line propagation

In a homogeneous, isotropic medium, nn is constant, so stationarity reduces to that of geometrical path length:

δ∫γds=0.\delta\int_\gamma ds=0.

In ordinary Euclidean space without additional constraints, the resulting ray is a straight line. This explains why light propagates rectilinearly in a uniform medium but may curve when the refractive index varies with position.

Law of reflection

Suppose a ray travels from AA, reflects at a point PP on a plane mirror, and reaches BB. In a uniform medium, its optical path length is

L=n(∣A−P∣+∣P−B∣).\mathcal{L}=n\bigl(|A-P|+|P-B|\bigr).

Varying PP along the mirror and requiring stationarity yields

θi=θr,\theta_i=\theta_r,

where the angles of incidence and reflection are measured from the surface normal. The incident ray, reflected ray, and normal lie in the same plane.

For a plane mirror, the construction can also be understood by reflecting one endpoint across the mirror: the broken reflected path then corresponds to a straight path to the reflected endpoint.

Snell’s law

For refraction at a plane interface between two homogeneous media,

L=n1ℓ1+n2ℓ2,\mathcal{L}=n_1\ell_1+n_2\ell_2,

where ℓ1\ell_1 and ℓ2\ell_2 are the distances traveled in the two media. Varying the crossing point along the interface gives Snell’s law:

n1sin⁡θ1=n2sin⁡θ2.n_1\sin\theta_1=n_2\sin\theta_2.

A ray entering a medium with a higher refractive index bends toward the normal. The stationary path balances geometrical distance against the different propagation speeds; it generally is not the shortest geometrical route.

When n1>n2n_1>n_2, a real transmitted ray exists only when

sin⁡θ1≤n2n1.\sin\theta_1\leq\frac{n_2}{n_1}.

Beyond the critical angle, geometrical transmission is replaced by total internal reflection. The electromagnetic field near the boundary still includes an evanescent field, which the elementary ray-path description does not represent.

Rays in continuously varying media

Fermat’s principle also describes propagation through a smoothly varying refractive index. Parameterizing a path by an arbitrary parameter uu gives

L=∫n(r)∣drdu∣ du.\mathcal{L}=\int n(\mathbf r) \left|\frac{d\mathbf r}{du}\right|\,du.

Applying the calculus of variations yields the ray equation

dds(ndrds)=∇n,\frac{d}{ds} \left(n\frac{d\mathbf r}{ds}\right) =\nabla n,

where ss is arc length. Writing the unit tangent as t=dr/ds\mathbf t=d\mathbf r/ds, this becomes

ndtds=∇n−t(t⋅∇n).n\frac{d\mathbf t}{ds} =\nabla n-\mathbf t(\mathbf t\cdot\nabla n).

Only the component of the refractive-index gradient perpendicular to the ray changes its direction. In an isotropic medium, rays therefore curve locally toward increasing refractive index.

This behavior underlies atmospheric mirages and graded-index optical systems. Variations in temperature and pressure can produce refractive-index gradients in the atmosphere; deliberately engineered gradients can guide light in lenses and optical fibers.

Connection with wave optics

Fermat’s principle belongs primarily to geometrical optics, the approximation in which wavelength is small compared with the characteristic scales of the optical system. Its physical basis can be understood through wave phase and interference.

For monochromatic light of vacuum wavelength λ0\lambda_0, the accumulated propagation phase along a path is

ϕ=2πλ0∫γn ds.\phi=\frac{2\pi}{\lambda_0}\int_\gamma n\,ds.

Nearby paths generally have different phases. In high-frequency wave descriptions, contributions from paths whose phase changes rapidly tend to cancel through interference, while neighborhoods of stationary-phase paths can contribute prominently. The stationary-phase condition is therefore equivalent to stationary optical path length.

This is an asymptotic connection, not a claim that light consciously selects a route or that a ray is an exact description at every scale.

The eikonal equation

A monochromatic scalar field can be approximated by

U(r)=A(r)eik0S(r),U(\mathbf r)=A(\mathbf r)e^{ik_0S(\mathbf r)},

where AA varies slowly, k0=2π/λ0k_0=2\pi/\lambda_0, and SS is the optical eikonal. At leading order in the short-wavelength approximation,

∣∇S∣2=n2.|\nabla S|^2=n^2.

This is the eikonal equation. In an isotropic medium, rays are normal to surfaces of constant SS, which are wavefronts.

The eikonal description can be derived from wave equations associated with Maxwell’s equations. It connects Fermat’s variational statement to the local evolution of optical phase.

Historical development

Ancient treatments of reflection already connected reflected paths with shortest-distance constructions. Heron of Alexandria is associated with a minimum-distance argument for reflection from a plane mirror.

In the seventeenth century, Pierre de Fermat extended this style of reasoning from distance to travel time. In 1662, he used a least-time principle to derive the law of refraction, taking light to travel more slowly in the optically denser medium.

The principle subsequently became part of a broader mathematical relationship between optics and mechanics. In classical mechanics, variational principles identify physical trajectories through stationarity of action. Optical path length plays an analogous role for rays. This correspondence was developed particularly through Hamilton’s work on optics and dynamics and helped establish the methods of Hamiltonian mechanics.

The modern term stationary optical path is more general and precise than the historical phrase least time.

Applications

Fermat’s principle is used to derive and analyze ray trajectories rather than merely to restate individual optical laws.

  • Optical design: Stationary optical paths determine how rays pass through combinations of lenses and mirrors. Equal optical path lengths from a point source to an ideal image point are associated with coherent focusing.
  • Ray tracing: Numerical methods calculate ray trajectories through piecewise-uniform or continuously varying media, using interface laws or differential ray equations.
  • Graded-index systems: Spatially controlled refractive-index profiles can bend rays without requiring a sequence of sharply curved interfaces.
  • Atmospheric propagation: Index variations explain curved ray paths, mirages, and changes in the apparent positions of distant objects.
  • Travel-time methods: Analogous stationary-time principles are used for acoustic and seismic rays, with wave speed replacing optical propagation speed.

In gravitational lensing, a generalized Fermat principle identifies images with stationary points of an arrival-time function. Such stationary points may be minima, maxima, or saddles. Its formulation involves curved spacetime and gravitational time delays, rather than simply a material refractive index.

Scope and limitations

Dispersion and pulse travel time

In a medium exhibiting dispersion, refractive index depends on frequency. At a fixed frequency, the usual optical-path formulation uses the phase refractive index:

δ∫γn(r,ω) ds=0.\delta\int_\gamma n(\mathbf r,\omega)\,ds=0.

The travel time of a narrowband pulse instead depends on group propagation. In a transparent isotropic medium, the group index is

ng=n+ω∂n∂ω.n_g=n+\omega\frac{\partial n}{\partial\omega}.

Consequently, stationary phase optical length must not automatically be interpreted as a minimum pulse travel time.

Diffraction and caustics

When apertures, obstacles, or material structures have dimensions comparable to the wavelength, diffraction becomes important. A single ray path cannot describe the full field.

At a caustic, neighboring rays converge or form an envelope, and elementary ray-amplitude formulas can become singular. Wave treatments remain necessary to calculate the finite field and its interference structure.

Anisotropic and more general media

The scalar expression ∫n ds\int n\,ds assumes isotropic propagation. In anisotropic materials, such as crystals exhibiting birefringence, propagation depends on direction and polarization. Energy-flow direction need not coincide with the wave normal, and generalized variational formulations are required.

Absorbing, moving, or time-dependent media also require additional care. The elementary fixed-endpoint expression alone does not determine attenuation, frequency changes, or the full electromagnetic behavior.

Finally, Fermat’s principle determines possible ray paths, not their intensities, phases upon reflection, polarization changes, or division of energy at an interface. Those properties require additional wave and electromagnetic relations.