Normal Modes

Strike a bell and it does not ring one note — it rings a chord. Pluck a guitar string anywhere, tap a bridge deck, kick the leg of a table, send an infrared photon through the atmosphere toward a CO $_2$ molecule: in every case the thing that vibrates back at you is not a single wiggle but a superposition of pure tones, each with its own frequency and its own frozen pattern of motion. Those pure tones are the normal modes, and they are the deepest organizing idea in all of vibration physics. A normal mode is a collective motion in which every part of the system oscillates at one common frequency, with fixed amplitude ratios — the pattern never changes shape, it only breathes in scale. A system with $N$ degrees of freedom has exactly $N$ of these special motions, and here is the miracle: every possible motion, however messy, is a sum of them. The modes are independent — each conserves its own energy, forever, never leaking to a neighbour — so an impossibly coupled $N$ -body problem collapses into a list of $N$ separate simple harmonic oscillators. Complexity in, arithmetic out. This article is about the concept: what a mode is, why fixed shape forces one shared frequency, how a struck system divides its energy among its modes, and what the modes of real systems look like — chains of masses, standing waves on a string, the three vibration patterns of CO $_2$ that make it a greenhouse gas, the swaying modes of skyscrapers. The concrete two-mass workhorse is solved in Coupled Oscillators; the systematic matrix machinery lives in The Normal-Mode Eigenvalue Problem. Here we learn to think in modes — the skill that later becomes phonons, photons, and the mode expansions of quantum field theory.

Key takeaways

A normal mode is a motion in which every coordinate of a coupled system oscillates sinusoidally at one shared frequency with fixed amplitude ratios: the shape is frozen, only its overall scale oscillates.
$N$ degrees of freedom → exactly $N$ normal modes, each with its own frequency $\omega_n$ and shape $\mathbf{a}^{(n)}$ .
The general motion is a superposition of all $N$ modes, $\mathbf{q}(t) = \sum_n \mathbf{a}^{(n)}(A_n\cos\omega_n t + B_n\sin\omega_n t)$ , with the $2N$ constants fixed by initial positions and velocities.
In a linear system each mode carries its own energy and never trades it with another mode — the modes are the coordinates in which the coupled system becomes $N$ independent simple harmonic oscillators.
Mode shapes of a uniform chain of beads are discrete sine patterns, $x_j \propto \sin(n\pi j/(N+1))$ , with frequencies $\omega_n = 2\sqrt{k/m}\,\sin\big(n\pi/2(N+1)\big)$ ; as $N \to \infty$ these become the standing waves of a string.
Real-world galleries: CO $_2$ 's symmetric stretch, antisymmetric stretch, and (doubly degenerate) bend; the sway modes of buildings; the vibrational modes every satellite is tested against.
Key idea: a normal mode is a fixed pattern that oscillates at one common frequency; the $N$ modes of an $N$ -coordinate system are independent, energy-conserving building blocks whose superposition is every motion the system can perform.

The core insight

A normal mode is a collective motion with one shared frequency and a frozen shape; a linear system with $N$ degrees of freedom has exactly $N$ of them, they oscillate independently — each keeping its own energy — and every possible motion is a superposition of modes with amplitudes and phases set by the initial conditions.

Where this lives today

Modal analysis is a daily industrial ritual. Every satellite and launch-vehicle payload is bolted to a shaker table and swept through its normal modes before flight, because a mode of the structure landing on a mode of the rocket is how payloads break; finite-element packages compute galleries of thousands of modes for aircraft wings, EV battery packs, and turbine blades as a matter of course. At the heaviest end, Taipei 101 carries a 660-tonne steel pendulum near its top — a tuned mass damper deliberately built as a second oscillator matched to the tower's lowest sway mode, soaking up its energy exactly by the coupled-resonant-transfer physics of this cluster of articles. The concept's reach now extends to the most exotic objects measured. When two black holes merge, the remnant rings down through its quasinormal modes — damped oscillations of spacetime itself with frequencies set by the final mass and spin; LIGO–Virgo–KAGRA fit these ringdown tones to test general relativity, a program literally called black-hole spectroscopy. And in quantum hardware, the normal modes of nanomechanical resonators and microwave cavities are cooled to their ground states and manipulated one phonon or photon at a time: each mode, being an independent harmonic oscillator, quantizes into its own ladder of energy levels. That last sentence is the door this article opens — a field, classical or quantum, is nothing but an infinite gallery of normal modes, and its quanta are the modes' excitations.

The physical picture

Watch a system move in a normal mode and the striking thing is its discipline. Every mass passes through equilibrium at the same instant; every mass reaches its extreme at the same instant; the ratios between the displacements never change. Freeze the motion at any moment and you always see the same photograph, merely rescaled. That photograph is the mode shape; the rate at which it breathes is the mode frequency. Nothing else about the motion is negotiable — start a system exactly in a mode shape, at rest, and it will oscillate in that shape at that frequency forever. Here is the full portrait gallery for three equal beads coupled by four equal springs between two walls — three degrees of freedom, hence exactly three modes:

Figure

The three normal modes of a three-bead chain — each pattern breathes in scale at its own frequency

Read the gallery from top to bottom and a universal trend emerges. The lowest mode is the most collective: all beads swing together, the springs between them barely stretch, and the restoring force per unit displacement is smallest — hence the gentlest frequency. Each successive mode adds one more sign change across the chain, one more "node," stretching the connecting springs harder and driving the frequency up. The highest mode is the most jagged — every neighbour opposed — and the stiffest. Counting nodes orders the spectrum, and this staircase of increasingly wrinkled patterns is exactly what will become the harmonics of a guitar string. Now the second half of the concept, the half that earns modes their keep. Start the system in anything other than a mode shape — strike one bead, say — and the motion looks hopelessly complicated: no bead traces a sinusoid, the pattern writhes, energy wanders from bead to bead. But underneath, the modes are simply running independently, each at its own frequency, and the mess is their sum:

Figure

A messy motion decomposed — what one mass does is exactly the sum of two clean mode sinusoids

The modes are the atoms of motion: indivisible, independent, and universal. Learning a system means learning its mode gallery — after that, every question about its dynamics is a question about a handful of independent simple harmonic oscillators.

What makes a motion a mode

Let us make the defining property precise. A system of $N$ coupled coordinates moves in a normal mode when every coordinate executes simple harmonic motion at one common frequency, with a fixed set of amplitude ratios and a common phase:

q_i(t) = a_i\,\cos(\omega t - \varphi), \qquad i = 1, \dots, N \qquad (1)

where: $q_i$ — the $i$ -th coordinate (a bead's displacement, a bond stretch, an angle); $a_i$ — the fixed amplitude of coordinate $i$ , whose ratios define the mode shape; $\omega$ — the single shared mode frequency; $\varphi$ — one common phase for the whole pattern. Two things are worth staring at. Every coordinate shares the same $\cos(\omega t - \varphi)$ : the pattern $(a_1, \dots, a_N)$ is multiplied by one global oscillating factor, which is exactly the statement that the shape only breathes in scale. And the shape is not arbitrary — only $N$ special shapes are self-consistent, one per mode. Why does a frozen shape force a single shared frequency? Because of how restoring forces work in a linear system, the two properties are actually the same property.

Derivation AOne frozen shape forces one shared frequency

We prove the defining equivalence of normal modes: if a linear system moves so that its shape never changes — all coordinates locked in fixed proportion — then every coordinate must oscillate sinusoidally at one common frequency, and the shape must be one the dynamics itself selects. The strategy is separation of variables in miniature: write the assumed motion as a fixed pattern times a single unknown function of time, feed it to the equations of motion, and watch the equations force that function to be a sinusoid. The step to watch is the "constant of separation" argument — a function of time cannot equal a different function of time unless both are constant — which is the same logic that later splits partial differential equations into space and time parts.

Step 1. Assume the shape is frozen: $q_i(t) = a_i f(t)$ for all $i$ , with fixed numbers $a_i$ and a single scale function $f(t)$ shared by every coordinate. Why this is the right formalization: "the pattern only breathes in scale" means precisely that all coordinates track one common time dependence, each scaled by its own fixed amplitude.

Step 2. Feed this into the linearized equations of motion. Near stable equilibrium every coordinate obeys an equation of the form $m_i \ddot{q}_i = -\sum_j K_{ij}\,q_j$ , where the constants $K_{ij}$ encode the springs and couplings. Substituting the ansatz and dividing by $m_i a_i f(t)$ :

\frac{\ddot{f}(t)}{f(t)} = -\,\frac{\sum_j K_{ij}\,a_j}{m_i\,a_i} \quad \text{for every } i

Why divide this way: the left side depends only on time, the right side only on the constants of the problem and the chosen shape — the equality must hold at every instant, for every coordinate.

Step 3. Apply the separation argument. The left side is one function of time; the right side is a constant (a different constant for each $i$ , potentially). For the equality to hold at all times, the left side must equal a single constant, call it $-\omega^2$ ; and then every right-hand side must equal that same constant. Why the constant is negative for stable systems: near a stable equilibrium the restoring forces push back toward the pattern's zero, making $\ddot{f}$ and $f$ have opposite signs.

Step 4. Read off both conclusions at once. From the time part, $\ddot{f} = -\omega^2 f$ , so $f(t) = \cos(\omega t - \varphi)$ : every coordinate oscillates sinusoidally at the same frequency $\omega$ — equation (1). From the shape part, $\sum_j K_{ij} a_j = \omega^2 m_i a_i$ for every $i$ : the shape cannot be arbitrary, but must reproduce itself under the action of the spring network, each coordinate's restoring force per unit inertia coming out in the same proportion $\omega^2$ . Why only special shapes survive: this self-consistency condition is an eigenvalue problem in disguise — for an $N$ -coordinate system it has exactly $N$ independent solutions, which is where the count "N degrees of freedom, N modes" comes from. The systematic solution method, $\det(K - \omega^2 M) = 0$ , is developed in The Normal-Mode Eigenvalue Problem. *For one coordinate the condition collapses to $k a = \omega^2 m a$ , i.e. $\omega = \sqrt{k/m}$ — the ordinary SHO; and any shape multiplied by a non-sinusoidal $f(t)$ fails Step 3, confirming that frozen shape and single shared frequency are inseparable.*

The simplest system with a genuine mode gallery is the two-mass, three-spring chain, worked out completely in Coupled Oscillators by the sum-and-difference trick. Quoting its result: the two shapes are the symmetric pattern $(1,1)$ , in which the coupling spring never stretches, and the antisymmetric pattern $(1,-1)$ , in which it is doubly engaged, with frequencies

\omega_1 = \sqrt{\frac{k}{m}} \;\; \text{for } (1,1), \qquad \omega_2 = \sqrt{\frac{k + 2\kappa}{m}} \;\; \text{for } (1,-1) \qquad (2)

where: $k$ — each wall spring; $\kappa$ — the coupling spring; $m$ — each mass. When all three springs are equal ( $\kappa = k$ ) the frequencies are $\sqrt{k/m}$ and $\sqrt{3k/m}$ . Note the pattern that generalizes: the collective, node-free shape is slow; the wrinkled shape that fights the coupling is fast — the same staircase as the three-bead gallery above.

Superposition: every motion is a chord

Modes would be a curiosity if systems only ever moved in one of them. Their real power is completeness. Because the equations of motion are linear, any sum of solutions is a solution — and it is a theorem of the eigenvalue machinery that the $N$ mode shapes span every possible configuration. So the most general motion of the system is a chord: all $N$ modes sounding at once, each with its own amplitude and phase,

\mathbf{q}(t) \;=\; \sum_{n=1}^{N} \mathbf{a}^{(n)}\,\big(A_n\cos\omega_n t + B_n\sin\omega_n t\big) \qquad (3)

where: $\mathbf{a}^{(n)}$ — the $n$ -th mode shape; $\omega_n$ — its frequency; $A_n, B_n$ — $2N$ constants in total, fixed by the $N$ initial positions and $N$ initial velocities. This is the classical mechanic's Fourier theorem: initial conditions in, mode amplitudes out, and from then on each term evolves in serene isolation. The construction of the special coordinates that make this decomposition automatic — and that diagonalize the energy — is the business of the Normal Coordinates article; the concept is what matters here: solving a linear vibrating system means finding its modes and dealing the initial conditions out to them. One consequence deserves its own paragraph, because it is the deepest fact in this article: the modes do not talk to each other. Each mode's amplitude, once set by the initial conditions, is set forever; each mode's energy is separately conserved. The complicated wandering of energy you see — from bead to bead, from one end of a struck structure to the other — is entirely an artifact of watching the wrong coordinates. In mode coordinates nothing wanders at all.

Go deeper · Derivation B · 2 minEach mode keeps its own energy

Derivation BEach mode keeps its own energy

We show, concretely on the two-mass system, that the total energy splits exactly into one energy per mode, with no cross terms, and that each mode's share is separately constant in time. The strategy is direct computation: write the kinetic and potential energies in terms of the mode amplitudes (the sum and difference coordinates of the 2-DOF system) and watch the cross terms cancel. The cancellation is not an accident of this example — it happens because the mode shapes are orthogonal in the appropriate sense, a property proved in general in The Normal-Mode Eigenvalue Problem — but seeing it happen by hand is what makes it believable. Watch the factor $\tfrac14$ : it comes from the change of variables $x_1 = (s+d)/2$ , and forgetting it breaks the energy bookkeeping.

Step 1. Set up the two-mass, three-spring system with coordinates $x_1, x_2$ and mode combinations $s = x_1 + x_2$ , $d = x_1 - x_2$ , so $x_1 = \tfrac12(s+d)$ , $x_2 = \tfrac12(s-d)$ . The energy is

E = \tfrac12 m\dot{x}_1^2 + \tfrac12 m\dot{x}_2^2 + \tfrac12 k x_1^2 + \tfrac12 k x_2^2 + \tfrac12 \kappa\,(x_1 - x_2)^2

Why the last term: only the coupling spring cares about the relative displacement; the wall springs each store energy in their own mass's displacement.

Step 2. Convert the kinetic energy. Squaring and adding, $\dot{x}_1^2 + \dot{x}_2^2 = \tfrac12(\dot{s}^2 + \dot{d}^2)$ — the cross term $\dot{s}\dot{d}$ appears with coefficient $+\tfrac12$ from $\dot{x}_1^2$ and $-\tfrac12$ from $\dot{x}_2^2$ and cancels. So $T = \tfrac14 m\dot{s}^2 + \tfrac14 m\dot{d}^2$ . Why the cross terms cancel: the shapes $(1,1)$ and $(1,-1)$ are orthogonal — the cancellation is the dot product $(1,1)\cdot(1,-1) = 0$ doing its work.

Step 3. Convert the potential energy the same way: $x_1^2 + x_2^2 = \tfrac12(s^2 + d^2)$ and $(x_1 - x_2)^2 = d^2$ , giving

V = \tfrac14 k\,s^2 + \tfrac14 (k + 2\kappa)\,d^2

Why the coupling lands only in $d$ : the symmetric mode never stretches the middle spring, so $\kappa$ cannot store energy in $s$ — the same physics that kept $\omega_1$ at the bare frequency.

Step 4. Collect the pieces. The energy is a clean sum of two single-oscillator energies with the correct mode frequencies from (2):

E = \underbrace{\tfrac14 m\big(\dot{s}^2 + \omega_1^2 s^2\big)}_{E_1} \;+\; \underbrace{\tfrac14 m\big(\dot{d}^2 + \omega_2^2 d^2\big)}_{E_2} \qquad (4)

where: $E_1, E_2$ — the mode energies. Since $s(t)$ and $d(t)$ each obey their own SHO equation, each bracket is the conserved energy of an independent oscillator: $E_1$ and $E_2$ are separately constant, for all time, whatever the initial conditions. The numeric example below checks $E_1 + E_2$ against the directly computed initial spring energy and gets exact agreement; and in the <a class="wiki-link" href="/wiki/aq-mech-coupled-oscillators" data-track="wiki_article.body.wiki_link">Coupled Oscillators</a> energy slosh, $E_1$ and $E_2$ stay constant even while the two masses trade all their energy — the trade is interference between modes, not flow between them.

The gallery: from bead chains to molecules to buildings

Chains of beads. Stretch the two-mass system to $N$ equal beads between two walls and the mode shapes organize themselves into a stunning pattern: the amplitude of bead $j$ in mode $n$ is a sampled sine curve, $a_j^{(n)} \propto \sin(n\pi j/(N+1))$ — the discrete sines visible in the three-bead gallery. The frequencies follow a closed formula worth deriving, because it contains the guitar string.

Go deeper · Derivation C · 2 minThe modes of a chain of N beads

Derivation CThe modes of a chain of N beads

We derive the complete mode spectrum of $N$ equal masses $m$ joined by equal springs $k$ between two walls: frequencies $\omega_n = 2\sqrt{k/m}\,\sin\big(n\pi/2(N+1)\big)$ with sine-curve shapes. The strategy is an educated guess with a self-consistency check — try a sampled sine wave as the shape and verify the equations of motion accept it. This works because a sine pattern has a magic property under second differences: shifting a sine left and right and averaging returns the same sine, rescaled. The step to watch is the boundary condition — it is the walls that quantize the allowed wavelengths, exactly as they will for the continuous string.

Step 1. Write the equation of motion of interior bead $j$ , pulled by its two neighbouring springs: $m\ddot{x}_j = k(x_{j+1} - x_j) - k(x_j - x_{j-1}) = k(x_{j+1} - 2x_j + x_{j-1})$ , with the wall conditions expressed as fictitious stationary beads $x_0 = x_{N+1} = 0$ . Why the combination $x_{j+1} - 2x_j + x_{j-1}$ : each spring pulls toward its far end, so the net force measures how much bead $j$ sits away from the average of its neighbours — a discrete second derivative.

Step 2. Guess the mode: $x_j(t) = \sin(j\theta)\cos\omega t$ for some angle $\theta$ per step. Use the identity $\sin\big((j{+}1)\theta\big) + \sin\big((j{-}1)\theta\big) = 2\sin(j\theta)\cos\theta$ :

-m\omega^2 \sin(j\theta) = k\,\big(2\cos\theta - 2\big)\sin(j\theta)

Why the guess survives: the sine pattern reproduces itself under the neighbour sum — every bead's equation reduces to the same scalar condition, which is precisely the self-consistency a mode requires.

Step 3. Solve the scalar condition using $1 - \cos\theta = 2\sin^2(\theta/2)$ :

\omega^2 = \frac{2k}{m}\,(1 - \cos\theta) = \frac{4k}{m}\,\sin^2\!\frac{\theta}{2} \quad\Longrightarrow\quad \omega = 2\sqrt{\frac{k}{m}}\,\Big|\sin\frac{\theta}{2}\Big|

Step 4. Impose the walls. $x_0 = 0$ holds automatically; $x_{N+1} = 0$ requires $\sin\big((N{+}1)\theta\big) = 0$ , i.e. $\theta = n\pi/(N+1)$ . The allowed modes are

\omega_n = 2\sqrt{\frac{k}{m}}\,\sin\!\Big(\frac{n\pi}{2(N+1)}\Big), \qquad n = 1, 2, \dots, N \qquad (5)

where: $n$ — the mode number, equal to the number of half-wavelengths of the sine pattern across the chain. Exactly $N$ distinct modes emerge: $n = N+1$ makes every $\sin(j\theta)$ vanish (no motion), and higher $n$ merely repeats earlier shapes with a sign — the count comes out right without being imposed. Note the ceiling: no mode of the chain, however wrinkled, exceeds $\omega_{\max} = 2\sqrt{k/m}$ , the frequency of neighbours in full opposition.

Step 5. Take the continuum limit. For $n \ll N$ , expand the sine: $\omega_n \approx \sqrt{k/m}\;n\pi/(N+1)$ . With bead spacing $a$ , chain length $L = (N+1)a$ , and wave speed $v = a\sqrt{k/m}$ , this is $\omega_n \approx n\pi v/L$ — in ordinary frequency,

f_n = \frac{n v}{2L} \qquad (6)

where: $v$ — the speed of waves on the effective string; $L$ — its length. These are exactly the harmonics of a string fixed at both ends: the fundamental and its integer multiples. The bead chain's low modes are the string's standing waves, sampled at $N$ points. Two quick checks confirm the formula: $N = 1$ gives $\omega = 2\sqrt{k/m}\sin(\pi/4) = \sqrt{2k/m}$ , correct for one mass between two springs; $N = 2$ gives $\omega/\sqrt{k/m} = 1$ and $\sqrt{3}$ , matching the equal-spring case of equation (2).

Figure

The continuum limit — beads on a chain already sit on the standing waves of a string

Molecules. A molecule of $n$ atoms has $3n$ degrees of freedom; removing overall translations and rotations leaves $3n - 6$ vibrational modes ( $3n - 5$ if the molecule is linear). CO $_2$ , linear and triatomic, has exactly four: the symmetric stretch (both oxygens breathe outward, carbon still) at about $1330\ \mathrm{cm^{-1}}$ , the antisymmetric stretch (oxygens shift one way, carbon the other) at $2349\ \mathrm{cm^{-1}}$ , and a bend at $667\ \mathrm{cm^{-1}}$ that counts twice, because the molecule can flex in two perpendicular planes — a degenerate pair. These are not abstractions: infrared light at $667$ and $2349\ \mathrm{cm^{-1}}$ is absorbed by exactly these modes, which is the molecular mechanism of the greenhouse effect. The symmetric stretch, leaving the molecule's dipole unchanged, is invisible to infrared and shows up in Raman scattering instead — mode shapes decide selection rules.

Figure

The four vibrational modes of carbon dioxide — mode shapes with their measured wavenumbers

Structures. A skyscraper, a bridge, a turbine blade, a satellite — each has its mode gallery, computed by finite-element software as a giant eigenvalue problem and measured by shaking the real object. A tall building's lowest modes are gentle side-to-side sways with periods of several seconds; higher modes twist and flex it. Engineers keep the mode frequencies away from earthquake spectra and wind-shedding frequencies — precisely because, as Resonance explains, driving any single mode near its frequency amplifies it by its quality factor.

Limiting cases

Case	Result	Meaning
$N = 1$	one mode, $\omega = \sqrt{k/m}$	a normal mode is simple harmonic motion; the whole theory reduces to the SHO
$N = 2$ , identical masses	shapes $(1,1)$ and $(1,-1)$ , eq. (2)	symmetry alone fixes the shapes — the fully worked case of Coupled Oscillators
$N \to \infty$ chain	$\omega_n \to n\pi v/L$ : string harmonics	discrete modes become standing waves; mechanics hands off to wave physics
unanchored system	one mode with $\omega = 0$	zero restoring force along free translation; the "mode" is uniform drift, not oscillation
symmetric system (e.g. CO $_2$ bend)	two or more modes share one frequency	degeneracy: any combination within the degenerate pair is also a mode
large amplitude	anharmonic terms couple the modes	mode independence is a linear privilege; energy begins to leak between modes

Common trap“Mode 2 is whatever mass 2 is doing”

A normal mode is not attached to a particular mass, spring, or atom — it is a collective pattern in which, in general, every part of the system moves. Mode 2 of the three-bead chain is not "bead 2's motion"; it is the pattern $(1, 0, -1)$ in which bead 2 happens to stand still while its neighbours oppose each other. The numbering orders the modes by frequency, not by geography. The same confusion in reverse is equally common: what a single mass does is not a mode — it is a mixture of all of them, which is why an individual struck bead traces the complicated, non-sinusoidal curve of the decomposition figure. Modes live in the space of configurations of the whole system; masses merely participate. If you can point at the moving part, you are probably looking at a coordinate, not a mode.

Worked examples

1 · Dealing a strike out to the modes (confidence builder)

Two $1.0\ \mathrm{kg}$ masses sit between walls with springs $k = 25\ \mathrm{N/m}$ at the walls and a coupling spring $\kappa = 12\ \mathrm{N/m}$ between them. Mass 1 is pulled aside $8.0\ \mathrm{cm}$ and both are released from rest. Find the mode frequencies, decompose the motion into modes, and compute each mode's energy. Does the motion ever repeat?

Step 1. Mode frequencies from (2): $\omega_1 = \sqrt{25/1.0} = 5.0\ \mathrm{rad/s}$ for the shape $(1,1)$ ; $\omega_2 = \sqrt{(25 + 24)/1.0} = \sqrt{49} = 7.0\ \mathrm{rad/s}$ for $(1,-1)$ .

Step 2. Decompose the initial state. With $s = x_1 + x_2$ and $d = x_1 - x_2$ : $s(0) = d(0) = 8.0\ \mathrm{cm}$ , all velocities zero. Both modes are excited with equal amplitude — a strike is always a chord.

Step 3. Write the motion from (3): $s(t) = 8.0\cos 5t$ and $d(t) = 8.0\cos 7t$ (in cm), so $x_1(t) = 4.0\cos 5t + 4.0\cos 7t$ and $x_2(t) = 4.0\cos 5t - 4.0\cos 7t$ . Neither mass moves sinusoidally, yet the system is just two pure tones playing at once.

Step 4. Mode energies from (4): $E_1 = \tfrac14 m\,\omega_1^2 s_0^2 = 0.25 \times 1.0 \times 25 \times (0.080)^2 = \boxed{0.040\ \mathrm{J}}$ and $E_2 = 0.25 \times 1.0 \times 49 \times (0.080)^2 = \boxed{0.078\ \mathrm{J}}$ . Cross-check against the directly computed initial energy: the stretched wall spring stores $\tfrac12(25)(0.080)^2 = 0.080\ \mathrm{J}$ and the coupling spring $\tfrac12(12)(0.080)^2 = 0.038\ \mathrm{J}$ , total $0.118\ \mathrm{J} = E_1 + E_2$ . ✓ Those two numbers, $0.040$ and $0.078\ \mathrm{J}$ , now remain constant forever — even while the masses trade energy back and forth every $\pi$ seconds.

Step 5. Periodicity check: the motion repeats when both $\cos 5t$ and $\cos 7t$ complete whole cycles, i.e. after $t = 2\pi\ \mathrm{s}$ (5 and 7 cycles respectively) — periodic, because $7/5$ is rational. Detune the springs so the ratio is irrational and the exact initial state never recurs.

2 · Weighing a chemical bond with a mode frequency (exam level)

Model CO $_2$ as three masses on a line — $m_{\mathrm{O}} = 16.0\,u$ , $m_{\mathrm{C}} = 12.0\,u$ — joined by two identical bond "springs" $k$ . The antisymmetric stretch is measured at $\tilde\nu_3 = 2349\ \mathrm{cm^{-1}}$ . Find the bond stiffness $k$ , predict the symmetric-stretch wavenumber, and compare with the measured $\approx 1333\ \mathrm{cm^{-1}}$ .

Step 1. Convert the wavenumber to an angular frequency: $\omega_3 = 2\pi c\,\tilde\nu_3 = 2\pi (2.998\times10^{10}\ \mathrm{cm/s})(2349\ \mathrm{cm^{-1}}) = 4.42\times10^{14}\ \mathrm{rad/s}$ .

Step 2. The mode frequencies of the linear three-mass model (standard results of the eigenvalue machinery) are $\omega_1 = \sqrt{k/m_{\mathrm{O}}}$ for the symmetric stretch — carbon sits still, so each oxygen oscillates on its own bond spring — and $\omega_3 = \sqrt{k\,(1/m_{\mathrm{O}} + 2/m_{\mathrm{C}})}$ for the antisymmetric stretch, where the light carbon's recoil against both springs raises the frequency.

Step 3. Solve for the bond stiffness. With $m_{\mathrm{O}} = 2.657\times10^{-26}\ \mathrm{kg}$ and $m_{\mathrm{C}} = 1.994\times10^{-26}\ \mathrm{kg}$ : $1/m_{\mathrm{O}} + 2/m_{\mathrm{C}} = 3.76\times10^{25} + 1.00\times10^{26} = 1.38\times10^{26}\ \mathrm{kg^{-1}}$ , so $k = \omega_3^2 / (1.38\times10^{26}) = (1.96\times10^{29})/(1.38\times10^{26}) = \boxed{1.4\times10^{3}\ \mathrm{N/m}}$ . A chemical double bond is in fact softer than a car's suspension spring by more than an order of magnitude; what makes it ring at $10^{14}\ \mathrm{rad/s}$ is not stiffness but the minute mass it has to move — measured with light.

Step 4. Predict the symmetric stretch: $\omega_3/\omega_1 = \sqrt{1 + 2m_{\mathrm{O}}/m_{\mathrm{C}}} = \sqrt{1 + 32/12} = 1.91$ , so $\tilde\nu_1 = 2349/1.91 = \boxed{1227\ \mathrm{cm^{-1}}}$ .

Step 5. Compare honestly: the measured value is $\approx 1333\ \mathrm{cm^{-1}}$ , so the two-spring model comes in about 8% low. The discrepancy is physics, not arithmetic — real CO $_2$ has a bond–bond interaction (stretching one bond stiffens the other) that a nearest-neighbour spring model omits, and the bending mode at $667\ \mathrm{cm^{-1}}$ needs a separate bending stiffness entirely. The model captures the mode structure exactly and the numbers to ten percent — and its failure pinpoints the next term in the molecular force field.

Origins · 1 minDaniel Bernoulli and how this was discovered

Daniel Bernoulli

1700–1782 · Basel and St Petersburg

Daniel Bernoulli was the first person to see that a complex vibration is a sum of simple ones. In 1733 he analyzed the hanging chain — a rope swinging from a fixed point — and found its family of oscillation shapes, each with its own frequency: the first mode gallery ever computed, complete with the function later named after Bessel. Twenty years later, in the great vibrating-string controversy with d'Alembert and Euler, Bernoulli staked out the claim that every motion of a string is a superposition of its sinusoidal modes — pure physical insight, since the mathematics to prove it (Fourier's) lay half a century in the future; his rivals' objections were sharpened into exactly the questions Fourier analysis eventually answered. Between these bookends he wrote Hydrodynamica (1738), founding fluid dynamics and the kinetic picture of gases. Modal superposition — the working language of everything from molecular spectroscopy to structural engineering — began as Bernoulli's minority opinion.

History

1733

— Daniel Bernoulli publishes the modes of the hanging chain: the first computed gallery of normal modes, with amplitude patterns given by what would later be called Bessel functions.

1747

— d'Alembert derives the wave equation for the vibrating string, framing the continuum problem the mode picture must explain.

1753

— Bernoulli asserts that every motion of a string is a superposition of its simple modes, igniting a decades-long dispute with Euler and d'Alembert about what functions a sine series can represent.

1822

— Fourier's Théorie analytique de la chaleur supplies the missing mathematics, turning Bernoulli's physical claim into the theorem underlying all mode decompositions.

1912

— Debye counts the normal modes of a crystal to predict its heat capacity — the step that, with quantization, turns classical modes into phonons.

Going further

References. Morin, Introduction to Classical Mechanics, Ch. 4 (Oscillations) — coupled oscillators and normal modes · Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., Ch. 6 (Oscillations) — eigenvalue formulation · Landau & Lifshitz, Mechanics, 3rd ed., §23–24 (Oscillations of systems with several degrees of freedom) · Taylor, Classical Mechanics, Ch. 11 (Coupled Oscillators and Normal Modes) The two-mass mode frequencies $\omega=\sqrt{k/m}$ and $\sqrt{3k/m}$ and the mass-chain dispersion $\omega_n=2\sqrt{k/m}\sin[n\pi/2(N+1)]$ are standard results; molecular mode counts $3n-6$ / $3n-5$ are textbook. Worked-example numbers (mode energies $0.040$ and $0.078$ J summing to the direct $0.118$ J; CO $_2$ bond constant $k \approx 1.4\times10^3$ N/m from $\tilde\nu_3 = 2349\ \mathrm{cm^{-1}}$ , predicted $\tilde\nu_1 = 1227$ vs measured $\approx 1333\ \mathrm{cm^{-1}}$ ) are arithmetic consequences of the boxed formulas; CO $_2$ wavenumbers are standard spectroscopic values.

Builds on: Simple Harmonic Motion · Small Oscillations About Equilibrium · Complex Exponential Representation of Oscillations · Amplitude, Phase, and Angular Frequency Leads to: Driven Oscillations · Resonance · Damped Harmonic Oscillator · Energy in Simple Harmonic Motion

Is a normal mode a motion, a shape, or a coordinate? The article seems to use all three.

All three faces belong to one object, used in different sentences. The motion is the full time history, equation (1): everything oscillating together at $\omega_n$ . The shape is the frozen amplitude pattern $\mathbf{a}^{(n)}$ — what you would photograph. And associated with each mode there is also a coordinate: the single number $\xi_n(t)$ measuring how much of the shape is currently present, which evolves like an independent SHO. Physicists slide between the three because the map is one-to-one: give the shape, and the frequency and coordinate come with it. The coordinate viewpoint — where the system's state is the list $(\xi_1, \dots, \xi_N)$ and the dynamics is $N$ separate oscillators — is developed properly in the Normal Coordinates article; this article's superposition formula (3) is exactly that list written back in ordinary coordinates.

Why exactly N modes for N degrees of freedom — not more, not fewer?

The self-consistency condition of the first derivation — the shape must reproduce itself under the spring network, $\sum_j K_{ij}a_j = \omega^2 m_i a_i$ — is an $N\times N$ eigenvalue problem, and such a problem has exactly $N$ independent solutions (counting degenerate ones separately). Intuitively: fewer than $N$ modes could not span all possible initial configurations, so some strikes would have no decomposition; more than $N$ would be redundant, since $N$ independent shapes already fill the configuration space. The bead chain shows the count working mechanically: sine patterns with $n = 1, \dots, N$ are all distinct, $n = N+1$ moves nothing, and higher $n$ repeats old shapes. The general proof — via $\det(K - \omega^2 M) = 0$ being degree $N$ in $\omega^2$ — lives in The Normal-Mode Eigenvalue Problem.

If every mode is perfectly periodic, why does the article say the general motion never repeats?

A sum of periodic functions is periodic only if their periods fit into a common one — that is, if all frequency ratios are rational. Two modes at $5$ and $7\ \mathrm{rad/s}$ recur together every $2\pi$ seconds (5 cycles meets 7 cycles), which is why the worked example repeats. But generic mode frequencies are incommensurate — $\omega_2/\omega_1$ irrational — and then the modes never simultaneously return to their starting phases: the motion comes arbitrarily close to its initial state infinitely often but never exactly repeats. This is not chaos — the motion is perfectly regular and predictable, just quasi-periodic. Real systems are effectively always in this case, which is why a struck bell's waveform, examined closely, never quite recurs even while each partial rings at its own steady frequency.

Where does mode independence actually break down in real systems?

At the same place the whole theory begins: the linear approximation. Modes decouple because the equations of motion are linear, which is the small-oscillation expansion of the true potential; push the amplitudes up and the neglected cubic and quartic terms act as weak couplings between modes, letting energy leak from one to another. The classic demonstration is the Fermi–Pasta–Ulam–Tsingou computation of 1955: energy fed into the lowest mode of a slightly nonlinear chain wandered among a few modes and then — famously, unexpectedly — almost entirely returned, instead of spreading evenly. That surprise seeded both chaos theory and soliton physics. In practice: molecular vibrations at high excitation exchange energy between modes (intramolecular energy redistribution), and structural engineers treat mode coupling as the signature that deformations have left the linear regime. Damping also couples what pure elasticity keeps separate.

What exactly is a zero-frequency mode, and is it really a mode?

Formally it solves the mode condition with $\omega = 0$ ; physically it is a direction of motion with no restoring force. Take the bead chain and remove the walls: shifting every bead equally stretches no spring, so the pattern $(1,1,\dots,1)$ feels zero force. Its "oscillation" equation is $\ddot{\xi} = 0$ , whose solution is uniform drift, $\xi(t) = \xi_0 + v_0 t$ , not a sinusoid — the free translation of the whole system. Every unanchored object carries such modes (three translations, three rotations for a molecule — exactly the $6$ subtracted in the $3n-6$ count). They matter practically: in any numerical mode computation, zero (or nearly zero) frequencies flag either genuine rigid-body freedom or a modeling error, and the degeneracy and zero-mode bookkeeping is handled systematically in The Normal-Mode Eigenvalue Problem.

Are mode shapes always sine curves, like in the bead chain?

No — sines are the special privilege of uniform systems. The chain's identical masses and springs make the discrete translation symmetry pick sinusoidal patterns, just as a uniform string gets $\sin(n\pi x/L)$ . Change anything — a heavier bead in the middle, a stiffer spring at one end, a tapered beam — and the mode shapes deform into system-specific patterns that only the eigenvalue machinery can find; Bernoulli's hanging chain, with gravity playing the role of position-dependent tension, has Bessel-function shapes. What survives in general is the structure: $N$ shapes, orthogonal in the mass-weighted sense, ordered by node count, each with its frequency. The sine-shaped case is still the one to memorize, because it is the bridge to Fourier series: for uniform systems, mode decomposition and Fourier decomposition are the same operation.

How do classical normal modes turn into phonons and field quanta?

Through the one fact this article kept emphasizing: each mode is an independent harmonic oscillator. Quantum mechanics knows exactly what to do with a harmonic oscillator of frequency $\omega_n$ — its energy comes in steps of $\hbar\omega_n$ . Quantizing a vibrating crystal therefore means quantizing each of its $3N$ modes separately: a state of the crystal is a list of occupation numbers, "how many quanta in each mode," and one quantum of a lattice mode is a phonon. The identical construction applied to the electromagnetic field's modes gives photons. Nothing about the mode concept changes — shapes, frequencies, and independence carry straight over; only the amplitude of each mode becomes a quantum ladder instead of a continuous dial. This is why normal modes are called the classical ancestor of field quanta, and why time invested here pays off all the way into quantum field theory.

Normal Modes

Where this lives today

The physical picture

What makes a motion a mode

Superposition: every motion is a chord

The gallery: from bead chains to molecules to buildings

Limiting cases

Worked examples

1 · Dealing a strike out to the modes (confidence builder)

2 · Weighing a chemical bond with a mode frequency (exam level)

Going further

Before this

Where this leads