The Talbot Effect: How a Grating Copies Itself With No Lens

What the Talbot Effect Is: The Physical Setup

The setup is deceptively simple. A periodic object — usually a one- or two-dimensional diffraction grating of period d — is illuminated by a coherent, collimated (plane-wave) beam, such as light from a laser. On a screen placed at the right distance behind the grating, an exact replica of the grating's transmission pattern appears, entirely without lenses or curved mirrors.

What makes this remarkable is that ordinary diffraction blurs a pattern as it propagates. Yet at special planes the diffracted waves re-conspire to rebuild the original. The phenomenon is near-field: it lives in the Fresnel diffraction regime, close enough to the grating that many diffraction orders still overlap, unlike the far-field Fraunhofer regime where you see only the grating's Fourier transform.

• Object: periodic grating (amplitude or phase), period d.
• Illumination: spatially coherent plane wave, wavelength λ.
• Output: self-images at multiples of the Talbot length z_T, and a rich interference field between them.

Henry Fox Talbot — better known as a pioneer of photography — first reported these "very curious" colored self-images in 1836.

The Mechanism: Why Diffraction Orders Rebuild the Grating

A periodic grating decomposes any transmitted field into a discrete set of diffraction orders. Each order n is a plane wave leaving at angle θ_n set by the grating equation sin θ_n = n·λ/d. As these plane waves propagate a distance z, each accumulates a phase that depends on its tilt.

The key is the longitudinal phase of order n: it goes as k_z,n = (2π/λ)·√(1 − (nλ/d)²). Expanding for small angles gives k_z,n ≈ (2π/λ)·[1 − (n²λ²)/(2d²)]. The n-dependent part is proportional to n². When z equals the Talbot length, every order has picked up a phase that is an integer multiple of 2π relative to the zeroth order — so all orders re-align exactly as they were at the grating, reconstructing the pattern.

• At z = z_T/2, the phase factor is (−1)^(n²), giving a self-image shifted by d/2.
• At rational fractions p/q of z_T, the orders combine into a superposition of shifted copies (a Gauss-sum structure), producing images with fractionally smaller periods.

Lord Rayleigh gave this constructive-interference explanation in 1881.

Key Quantities and a Worked Example

The exact Talbot length is z_T = λ / (1 − √(1 − λ²/d²)). For the common case d ≫ λ this simplifies to the textbook form:

z_T ≈ 2d² / λ.

Here d is the grating period, λ the wavelength. (A widely used alternative convention writes z_T = d²/λ; the factor-of-2 difference is just whether you count the half-period-shifted image at d²/λ as a "Talbot plane.") Note the strong scaling: halving the period quarters the distance.

• Worked example: d = 20 µm, λ = 633 nm (HeNe red). Then z_T ≈ 2·(20×10⁻⁶)² / (633×10⁻⁹) ≈ 2·(4×10⁻¹⁰)/(6.33×10⁻⁷) ≈ 1.26×10⁻³·... → z_T ≈ 1.26 mm. The first (half-shifted) image at z_T/2 sits at ~0.63 mm.
• For coarse gratings (d = 100 µm) this stretches to z_T ≈ 3.2 cm; for a 2 µm period it collapses to z_T ≈ 12.6 µm.
• When λ approaches d, the exact formula is mandatory — the 2d²/λ approximation can err by up to 100%.

How It's Observed, Measured, and Applied

Observing the Talbot effect requires only spatial coherence and a periodic object. In the lab you illuminate a Ronchi ruling with an expanded HeNe or diode laser and translate a camera or screen along the beam; self-images blink into sharp focus at each z_T interval, with the intermediate Talbot carpet visible as a woven interference pattern.

Applications exploit the lensless, self-repeating imaging:

• Talbot-Lau interferometry: uses two or three gratings to sense phase shifts; the basis of X-ray phase-contrast imaging, which reveals soft tissue that absorbs X-rays too weakly for conventional radiography.
• Displacement and alignment sensing: lateral shifts of a self-image measure sub-micron motion; used in encoders and wavefront sensing.
• Array illumination: fractional Talbot planes multiply a grating into denser spot arrays for optical computing and lithography.
• Spectroscopy: because z_T depends on λ, white light produces color-separated self-images — exactly Talbot's original 1836 observation.

The effect also appears in photonic-integrated multimode interference couplers, which rely on the same self-imaging in a waveguide.

Comparison to Related Diffraction Regimes

The Talbot effect is often confused with its neighbors, so it helps to place it in the diffraction landscape:

• Fresnel (near-field) diffraction: the Talbot effect is a special, periodic case of Fresnel diffraction. Non-periodic apertures do not self-image; periodicity is what forces the orders to re-phase collectively.
• Fraunhofer (far-field) diffraction: at very large z you see the grating's discrete diffraction spots — the Fourier transform — not a real-space copy. The Talbot regime ends when only one order survives overlap.
• Lens imaging: a lens forms a single conjugate image at a location set by 1/v = 1/f − 1/u. Talbot self-imaging is lensless, periodic in z, and confined to periodic objects.
• Moiré / shadow effects: pure geometric shadow of a grating has no revival planes; the Talbot revivals are genuinely wave-interferometric.

The same mathematics — quadratic phase accumulation of Fourier components — reappears as the temporal Talbot effect in dispersive fibers, where a periodic pulse train self-images in time rather than space.

Significance, Universality, and Open Questions

The Talbot effect's deepest lesson is universality: any wave with a quadratic dispersion relation and a periodic source will self-image. It has therefore been demonstrated far beyond visible light.

• Matter waves: electron Talbot patterns (Chapman et al., 1995) and atomic self-imaging confirmed de Broglie waves obey the same law, with z_T = 2d²/λ_dB and λ_dB = h/(mv) as small as picometers.
• X-rays, plasmons, water waves, and spin waves all show Talbot revivals.
• Quantum and nonlinear regimes: the fractional Talbot effect is linked to Gauss sums from number theory, and "Talbot carpets" have been proposed as physical factorizers of integers. A nonlinear Talbot effect (2011, in optical superlattices) self-images the second-harmonic field.

A famous subtlety: near a phase-grating, the field can exhibit Talbot fractals — self-similar structure at every scale — studied by Berry and Klein (1996), who named the intensity landscape the "Talbot carpet." Open questions include how coherence, source size, and grating imperfections blur revivals, and how to harness the temporal Talbot effect for optical-clock pulse-rate multiplication.

Talbot self-images at fractional distances behind a grating of period d (amplitude grating, plane-wave illumination)

Distance z	What appears	Effective period	Lateral shift
z = z_T	Exact self-image	d	0
z = z_T / 2	Self-image (contrast inverted / half-shifted)	d	d/2
z = z_T / 4	Image with half the period	d/2	d/4
z = z_T / 8	Image with quarter period (fractal sub-images)	d/4	d/8
z = p/q · z_T	Superposition of q shifted copies	d/q (q odd)	varies