Chladni Patterns

The intuition — sand finds the places that don't move

Clamp a thin metal plate at its center, scatter fine sand across it, then draw a violin bow down one edge. At most frequencies nothing interesting happens — the sand just dances chaotically. But hit a resonant frequency and, within a second or two, the chaos resolves into a crisp geometric figure: crosses, stars, concentric rings, checkerboards.

The trick is simple once you see it. When the plate resonates it vibrates in a standing-wave mode. Some regions swing up and down with large amplitude — the antinodes. Other curves stay perfectly still — the nodal lines. A grain of sand sitting on an antinode gets flung into the air every cycle; when it lands it has drifted a little, doing a tiny random walk. A grain that happens to land on a nodal line feels no vertical kick, so it stops moving. Over thousands of cycles the sand drains off every moving region and accumulates on the still lines. The pattern you see is a direct photograph of the mode's nodal set.

This is why a single plate produces dozens of different figures: each resonant frequency excites a different mode, and each mode has its own arrangement of nodal lines.

How it works — vertical acceleration sorts the grains

The sorting mechanism is the plate's vertical acceleration, not its displacement. At an antinode the surface moves as w(t) = A·sin(ωt), so its acceleration is −Aω²·sin(ωt). When that downward acceleration exceeds g (≈ 9.81 m/s²), the plate momentarily falls away faster than the grain can — the grain goes ballistic and leaves the surface. A 200 Hz mode with just 6 µm of amplitude already produces a peak acceleration of A·ω² = (6×10⁻⁶)(2π·200)² ≈ 9.5 m/s², right at the launch threshold, so even invisibly small vibrations toss the sand.

Each hop lands the grain a short distance away, biased — on average — toward regions of lower acceleration, because grains spend longer airborne where they were thrown hardest. The only stable resting places are the nodal lines, where the acceleration is identically zero. The result is a self-cleaning effect: the antinodes scrub themselves clean and feed the nodes.

The governing physics — the biharmonic plate equation

A thin, flat, elastic plate doesn't obey the ordinary wave equation that governs a drum membrane or a guitar string. A membrane resists deflection through tension; a plate resists through bending stiffness. That changes the spatial derivative from second order to fourth order. The Kirchhoff–Love thin-plate equation for free, undamped transverse vibration is:

D ∇⁴w  +  ρh ∂²w/∂t²  =  0

where w(x, y, t) is the out-of-plane displacement, ρ is the material density, h is the plate thickness, and D is the flexural rigidity:

D = E·h³ / [ 12 (1 − ν²) ]

Here E is Young's modulus and ν is Poisson's ratio. The operator ∇⁴ is the biharmonic operator:

∇⁴w = (∂²/∂x² + ∂²/∂y²)² w
     = ∂⁴w/∂x⁴ + 2 ∂⁴w/∂x²∂y² + ∂⁴w/∂y⁴

Look for a standing-wave solution w(x, y, t) = W(x, y)·cos(ωt). Substituting turns the time-dependent equation into the eigenvalue problem:

∇⁴W = (ρh / D) ω² W   →   ∇⁴W = k⁴ W,   with   k⁴ = ρh ω² / D

The mode shape W(x, y) is the eigenfunction; its zero set (the curves where W = 0) is exactly the nodal pattern the sand reveals. The boundary conditions — on a classic Chladni plate the edges are free (no shear, no bending moment) — quantize the problem into a discrete spectrum of frequencies ω₁ < ω₂ < ω₃ < … and their mode shapes.

Why plates scale as frequency² — the dispersion relation

The fourth-order equation gives plates a distinctive signature. For a travelling bending wave of wavenumber k the dispersion relation is:

ω = k² · √(D / ρh)

Frequency grows with the square of the wavenumber. Contrast a string, where ω = v·k grows linearly. The consequences are concrete:

• Bending waves are dispersive. High-frequency components travel faster than low-frequency ones (phase speed v = ω/k = k·√(D/ρh) rises with frequency). A sharp tap on a plate spreads into a "chirp."
• Higher modes crowd together. Doubling the number of nodal lines roughly quadruples the frequency, so the spectrum of plate modes climbs steeply. A square steel plate might show a half-dozen distinct figures between 200 Hz and 2 kHz.
• Chladni's law follows. For a circular plate the empirical fit Chladni published is f ≈ C(m + 2n)², where m counts diametric nodal lines and n counts circular ones — the square dependence is the fingerprint of the biharmonic operator.

Mode shapes on a square plate

For a center-supported square plate of side L the modes are labelled by two integers (m, n) counting the half-wavelengths along each edge. A useful conceptual approximation treats each mode as a product (or symmetric combination) of one-dimensional shapes, with a frequency that rises with m² + n². Degenerate modes (where two different (m, n) pairs share the same frequency, e.g. (2,1) and (1,2)) combine into rotated or diagonal figures — which is why the same plate can show both a plus-sign and an X at nearby frequencies.

Mode (m, n)	Nodal figure on a square plate	Relative frequency (∝ m²+n²)
(1, 1)	Single diagonal cross / saddle	2
(2, 1) & (1, 2)	Plus-sign or X (degenerate pair)	5
(2, 2)	2×2 checkerboard, 4 cells	8
(3, 1) & (1, 3)	Three parallel bands, rotatable	10
(3, 2) & (2, 3)	Lattice of 6 cells	13
(3, 3)	3×3 grid, 9 cells	18
(4, 4)	4×4 grid, dense star-lattice	32

Real plates deviate from this idealized table because free edges, the central support, anisotropy of the metal, and small thickness variations all shift and split the frequencies. But the qualitative march — more nodal lines as frequency climbs — is exactly what you watch happen on the plate.

Numbers — what a real plate does

Consider a steel plate 300 mm square and 1.0 mm thick (E ≈ 200 GPa, ρ ≈ 7850 kg/m³, ν ≈ 0.3). Its flexural rigidity is D = Eh³/[12(1−ν²)] ≈ 18.3 N·m, and √(D/ρh) ≈ 1.53 m²/s.

Quantity	Symbol	Value (steel example)
Flexural rigidity	D = Eh³/12(1−ν²)	≈ 18.3 N·m
Bending-wave speed factor	√(D/ρh)	≈ 1.53 m²/s
Lowest audible mode	f₁	roughly 100–300 Hz
Mode spacing (low modes)	Δf	tens to ~100 Hz
Amplitude needed to throw sand	A at 200 Hz, a = g	≈ 6 µm
Peak acceleration at antinode	Aω²	up to many ×g
Grains per visible line	—	10³–10⁵ depending on dosing

The headline number worth remembering: at a few hundred hertz, a vibration amplitude of only a few microns — far too small to see — already exceeds the 1 g acceleration that launches the sand. The pattern makes an invisible motion field visible by amplifying it through thousands of bounce cycles.

Where Chladni patterns show up

• Musical instrument making. Violin, viola, cello, and guitar luthiers dust the unassembled top and back plates and excite them to read the nodal figures (the "ring modes" or modes #2 and #5). The shape and frequency of these modes guide where to shave wood, tuning the plate before the box is glued together.
• Structural and mechanical engineering. Loudspeaker cones, hard-disk platters, brake rotors, circuit boards, solar-panel substrates, and aircraft skin panels all have plate modes. Engineers locate the nodal lines (now via finite-element modal analysis, confirmed by laser-Doppler vibrometry or electronic speckle-pattern interferometry) to know where mounting points, stiffeners, or damping pads do the most good.
• Turbomachinery. Turbine and compressor blades are essentially tapered plates; their nodal diameters and circles are mapped during design because a blade resonating at engine-order frequencies fails by fatigue. The classic dusting demo is the conceptual ancestor of modern blade modal testing.
• Acoustic particle manipulation. The same node-trapping physics, run with ultrasound in a fluid, sorts cells and microparticles in lab-on-chip "acoustofluidic" devices and assembles particle arrays at pressure nodes.
• Education and art (cymatics). The plate-and-sand demo is a staple of physics teaching, and artists use sound-driven plates and membranes to make the geometry of vibration visible.

A two-century-old experiment

Robert Hooke ran an early version in 1680, bowing a glass plate dusted with flour. Ernst Chladni systematized it in 1787 in Entdeckungen über die Theorie des Klanges ("Discoveries in the Theory of Sound"), cataloguing the figures and touring Europe with the demonstration. Napoleon was reportedly impressed enough to fund a prize for explaining the figures mathematically; Sophie Germain won it in 1816 with the first plate theory, though her derivation still had errors (Lagrange had already corrected the fourth-order equation itself in 1811). The fully consistent thin-plate theory — with the correct free-edge boundary conditions — came from Gustav Kirchhoff in 1850. Michael Faraday added a crucial twist in 1831, showing that fine powders form the inverse figures because air currents over the plate carry light particles to the antinodes.

Common misconceptions and edge cases

• "The sand marks where the plate vibrates most." Exactly backwards. Sand marks where the plate is still — the nodal lines. The blank regions are the antinodes that vibrate hardest.
• "It's the same as standing waves on a string or drum." Only loosely. Strings and drum membranes obey a second-order wave equation (frequency linear in mode index, ω ∝ k). A plate's restoring force is bending stiffness, giving a fourth-order biharmonic equation and ω ∝ k². That's why plate patterns are geometrically richer and the spectrum climbs faster.
• "Wet sand should look the same, just heavier." Faraday's surprise: fine or damp powder can pile on the antinodes instead, driven by acoustic streaming (the steady air current set up over the oscillating surface) and capillary clumping. Particle size and dampness flip the figure.
• "Free edges and clamped edges give the same pattern." No — boundary conditions change the eigenfunctions and the frequencies. A free-edge square plate, a clamped circular plate, and a center-supported plate each have distinct mode families and nodal sets at the same physical size.
• "You hear the resonance, so it's loudest at resonance." The visible figure forms because the response is sharply amplified at resonance (high Q), but a damped plate broadens and weakens the figures; very lossy materials (rubber, soft plastics) barely pattern at all because the antinodes never reach the launch acceleration.
• "Higher frequency always means more lines." Generally yes, but degeneracies and mode crossings mean two nearby frequencies can show very different figures, and small frequency changes can rotate or merge a pattern rather than add lines.