Electromagnetism
The Larmor Radiation Formula
Any accelerating charge radiates — and the power goes as acceleration squared: P = q²a²/(6πε₀c³)
The Larmor formula gives the total electromagnetic power radiated by a non-relativistic accelerating point charge: P = q²a²/(6πε₀c³), where q is the charge (coulombs), a is the instantaneous acceleration (m/s²), ε₀ = 8.854 × 10⁻¹² F/m is the vacuum permittivity, and c = 2.998 × 10⁸ m/s is the speed of light. Derived by Joseph Larmor in 1897, it says radiated power scales with the square of the acceleration — a charge at constant velocity radiates nothing, while an oscillating or orbiting charge continuously loses energy. It is the reason a classical orbiting electron should spiral into the nucleus, the origin of the sin²θ dipole pattern, and the parent of both synchrotron radiation and bremsstrahlung.
- Larmor formula (SI)P = q²a² / (6πε₀c³)
- Gaussian unitsP = 2q²a² / (3c³)
- Angular patterndP/dΩ ∝ sin²θ (dipole torus)
- ScalingP ∝ a² (independent of velocity, v≪c)
- Discoverer / yearJoseph Larmor, 1897
- Relativistic formP = q²γ⁶[a²−(v×a/c)²]/(6πε₀c³)
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
The formula, symbol by symbol
The Larmor formula is the total power (energy per unit time) carried away as electromagnetic radiation by a single point charge whose speed is small compared to light:
P = q² a² / (6 π ε₀ c³)
- P — total radiated power, watts (W = J/s).
- q — the charge, coulombs (C). For an electron, q = −1.602 × 10⁻¹⁹ C; only q² appears, so sign is irrelevant.
- a — magnitude of the instantaneous acceleration, m/s². It is the square of this quantity that governs the output.
- ε₀ — permittivity of free space, 8.854 × 10⁻¹² F/m (equivalently C²·N⁻¹·m⁻²).
- c — speed of light in vacuum, 2.998 × 10⁸ m/s. Because it enters as c³, radiation is a weak effect at everyday accelerations.
In Gaussian (CGS) units, where the awkward 4πε₀ disappears, the same physics reads P = 2q²a²/(3c³). The two forms are identical once you substitute the unit conventions; the SI version simply makes ε₀ explicit.
Why the power goes as acceleration squared
Start from the fields of a moving charge. Far from the charge, the radiation field (also called the acceleration field) dominates. For v ≪ c it has magnitude
E_rad = q·a·sinθ / (4πε₀ c² r)
Two features are decisive. First, E_rad is proportional to the acceleration a — no acceleration, no radiation field. Second, it falls off only as 1/r, not the 1/r² of a static Coulomb field. The Poynting vector — the energy flux — is proportional to E², so it goes as a²/r². When you integrate that flux over a sphere of radius r, whose area grows as r², the r-dependence cancels exactly, leaving a finite total power that depends on r not at all and on a² directly. That surviving 1/r radiation tail is precisely what allows energy to escape to infinity.
Doing the angular integral of the sin²θ pattern (below) supplies the numerical factor 6π, and the result is the Larmor formula. The physical takeaway: uniform motion is silent, acceleration is loud, and the loudness grows as the square of how hard you shake the charge.
The sin²θ dipole radiation pattern
Radiation is not emitted equally in all directions. The power per unit solid angle is
dP/dΩ = q² a² sin²θ / (16 π² ε₀ c³)
where θ is measured from the acceleration vector. The consequences:
- At θ = 90° (perpendicular to the acceleration) the emission is maximal.
- At θ = 0° and 180° (straight along the acceleration) the emission is exactly zero — a charge does not radiate in the direction it is being pushed.
- The three-dimensional pattern is a torus (doughnut) wrapped around the acceleration axis — the very same shape produced by a short dipole antenna, which is just a charge sloshing back and forth.
Integrating sin²θ over the full sphere gives ∫sin²θ dΩ = 8π/3, which converts dP/dΩ into the total P = q²a²/(6πε₀c³). This is why the effect is called dipole radiation: the leading term in the multipole expansion of any accelerating charge distribution is the electric dipole, and its angular signature is sin²θ.
The classical atom should collapse
Here is the historically explosive consequence. Model the hydrogen atom classically: an electron in a circular orbit of radius r₀ = 5.29 × 10⁻¹¹ m (the Bohr radius) around a proton. Circular motion means the electron is always accelerating toward the center with
a = v²/r = e² / (4πε₀ m_e r²) ≈ 9 × 10²² m/s²
Plug that into Larmor. The orbiting electron continuously radiates power on the order of 10⁻⁸ W, bleeding away its kinetic and potential energy. As it loses energy it spirals inward, accelerating harder, radiating even more — a runaway collapse. Integrating the energy loss gives a lifetime of only about
t_collapse ≈ 1.6 × 10⁻¹¹ s
In other words, a classical atom would emit a burst of ultraviolet light and self-destruct in about 16 picoseconds. Matter could not exist. The blatant contradiction with a stable, structured universe was one of the sharpest failures of classical physics and a direct spur toward Bohr's 1913 quantized orbits and, ultimately, quantum mechanics — where stationary states have definite energy and simply do not radiate.
Worked example: an electron in an antenna
Suppose an electron oscillates sinusoidally, x(t) = x₀ cos(ωt), with amplitude x₀ = 1 nm at frequency f = 10 GHz (ω = 2πf ≈ 6.28 × 10¹⁰ rad/s). The peak acceleration is a₀ = ω²x₀ ≈ (6.28 × 10¹⁰)² × 10⁻⁹ ≈ 3.9 × 10¹² m/s². The time-averaged radiated power (using ⟨a²⟩ = a₀²/2) is
⟨P⟩ = q² a₀² / (12 π ε₀ c³)
= (1.6×10⁻¹⁹)² (3.9×10¹²)² / (12π · 8.854×10⁻¹² · (3×10⁸)³)
≈ 4 × 10⁻²⁹ W
Tiny for one electron — but multiply by the ~10²³ electrons oscillating coherently in a real antenna and the numbers become the milliwatts and watts of everyday radio. This is the microscopic root of all antenna emission: coherent Larmor radiation from many charges.
Synchrotron radiation and bremsstrahlung
Any acceleration radiates, but the direction of the acceleration relative to the velocity names the phenomenon:
| Mechanism | Source of acceleration | Geometry | Typical setting |
|---|---|---|---|
| Synchrotron / cyclotron radiation | Magnetic Lorentz force (transverse) | a ⟂ v (centripetal) | Storage rings, pulsars, radio galaxies |
| Bremsstrahlung ("braking radiation") | Coulomb field of a nucleus | a mostly ∥ v (deceleration) | X-ray tubes, electron-target collisions |
| Dipole / antenna radiation | Applied oscillating field | a oscillating along wire | Radio, TV, cellular transmitters |
| Thomson / Compton scattering | Incident EM wave drives the charge | a set by wave polarization | Light scattering, plasma diagnostics |
Both synchrotron and bremsstrahlung are computed from the relativistic Liénard generalization of Larmor,
P = q² γ⁶ [ a² − (v × a / c)² ] / (6 π ε₀ c³)
where γ = 1/√(1 − v²/c²) is the Lorentz factor. For a particle on a circular path (a ⟂ v) this reduces to a γ⁴ enhancement; for straight-line acceleration it reduces to γ⁶. At the γ ≈ 10⁴ of a modern electron storage ring, this factor is what makes synchrotron light sources billions of times brighter than a laboratory X-ray tube, and it is why light electrons radiate far more than heavy protons at the same energy — the acceleration for a given force scales as 1/mass, so P scales as 1/mass².
Larmor radiation in numbers
| Situation | Acceleration a (m/s²) | Order-of-magnitude radiated power |
|---|---|---|
| Electron in Bohr ground-state orbit | ≈ 9 × 10²² | ~10⁻⁸ W (→ 16 ps collapse) |
| Electron, 10 GHz, 1 nm amplitude | ≈ 3.9 × 10¹² | ~4 × 10⁻²⁹ W (per electron, time-avg) |
| Proton vs electron, same force | a_p ≈ a_e / 1836 | P_p ≈ P_e / 1836² (~3×10⁻⁷) |
| Charge at constant velocity | 0 | 0 W (no radiation) |
| Electron in a synchrotron (γ≈10⁴) | huge transverse a | γ⁴-enhanced — kilowatts per beam |
JavaScript — Larmor power calculations
const q_e = 1.602176634e-19; // elementary charge, C
const eps0 = 8.8541878128e-12; // vacuum permittivity, F/m
const c = 2.99792458e8; // speed of light, m/s
const m_e = 9.1093837015e-31; // electron mass, kg
// Larmor formula (non-relativistic): total radiated power
function larmorPower(q, a) {
return (q * q * a * a) / (6 * Math.PI * eps0 * c * c * c);
}
// Bohr ground-state orbit: centripetal acceleration a = v^2 / r
const a0 = 5.29177e-11; // Bohr radius, m
const v_orbit = q_e / Math.sqrt(4 * Math.PI * eps0 * m_e * a0); // ~2.19e6 m/s
const a_centripetal = v_orbit * v_orbit / a0; // ~9e22 m/s^2
console.log(`Orbit acceleration: ${a_centripetal.toExponential(2)} m/s^2`); // ~9.0e22
console.log(`Radiated power: ${larmorPower(q_e, a_centripetal).toExponential(2)} W`); // ~4.6e-8
// Angular distribution: power per unit solid angle at angle theta from a-vector
function dPdOmega(q, a, thetaRad) {
const s = Math.sin(thetaRad);
return (q * q * a * a * s * s) / (16 * Math.PI * Math.PI * eps0 * c * c * c);
}
console.log(`dP/dOmega at 90 deg: ${dPdOmega(q_e, a_centripetal, Math.PI/2).toExponential(2)}`); // max
console.log(`dP/dOmega at 0 deg: ${dPdOmega(q_e, a_centripetal, 0).toExponential(2)}`); // 0
// Relativistic (Lienard) generalization for circular motion (a perpendicular v)
function lienardCircular(q, a, gamma) {
return larmorPower(q, a) * Math.pow(gamma, 4);
}
console.log(`gamma=1e4 circular: ${lienardCircular(q_e, 1e18, 1e4).toExponential(2)} W`);
// Time-averaged power of a sinusoidal oscillator, x = x0 cos(w t)
function oscillatorAvgPower(q, x0, omega) {
const aPeak = omega * omega * x0; // peak acceleration
return larmorPower(q, aPeak) / 2; // = aPeak^2 / 2
}
const omega = 2 * Math.PI * 10e9; // 10 GHz
console.log(`10 GHz e- osc: ${oscillatorAvgPower(q_e, 1e-9, omega).toExponential(2)} W`); // ~4e-29
Common misconceptions
- "A charge radiates because it is moving." No — only accelerating charges radiate. A charge in uniform motion (constant velocity) has a = 0 and emits nothing; its field is just a Lorentz-boosted Coulomb field carrying no energy to infinity.
- "The radiation is strongest along the direction of acceleration." Exactly backwards. The sin²θ pattern is zero along the acceleration axis and maximal perpendicular to it.
- "Larmor works at any speed." Only for v ≪ c. At relativistic speeds you must use the Liénard formula with its γ⁶ (or γ⁴ for circular motion) factors; ignoring them underestimates synchrotron output by many orders of magnitude.
- "Heavy and light particles radiate the same for the same force." For a given force, a = F/m, so radiated power scales as 1/m². An electron radiates ~1836² ≈ 3.4 million times more than a proton pushed by the same force — the reason electron machines, not proton machines, dominate synchrotron light production.
- "Larmor forbids stable atoms, so it must be wrong." It is not wrong classically; it correctly shows the classical atom is unstable. The fix is quantum mechanics, not a repair of Larmor. The formula remains exact for classical, non-relativistic point charges.
- "P depends on velocity." In the non-relativistic Larmor formula it does not appear at all — only a² does. Velocity re-enters only through the relativistic γ factors.
Frequently asked questions
What is the Larmor formula?
The Larmor formula gives the total electromagnetic power radiated by a non-relativistic accelerating point charge: P = q²a²/(6πε₀c³), where q is the charge, a is the magnitude of the instantaneous acceleration, ε₀ = 8.854 × 10⁻¹² F/m is the vacuum permittivity, and c = 2.998 × 10⁸ m/s is the speed of light. In Gaussian units it reads P = 2q²a²/(3c³). It was derived by Joseph Larmor in 1897 and is a cornerstone of classical electrodynamics.
Why is radiated power proportional to acceleration squared?
The radiation (far) field of a point charge is proportional to its acceleration, E_rad ∝ q·a·sinθ/r, falling off as 1/r rather than 1/r². The Poynting flux — energy per area per second — is proportional to E², so it scales as a²/r². Because that flux falls off only as 1/r², integrating it over a sphere of radius r (area ∝ r²) leaves a finite, r-independent total power proportional to a². A charge moving at constant velocity has zero acceleration and radiates nothing.
Why don't electrons in atoms spiral into the nucleus?
Classically they should. An electron orbiting a proton has centripetal acceleration a = v²/r ≈ 9 × 10²² m/s² in the ground-state Bohr orbit, so the Larmor formula predicts it radiates energy and spirals in within about 1.6 × 10⁻¹¹ seconds — the classical atom is unstable. The resolution is quantum mechanics: stationary states have definite, quantized energy and do not radiate. The Larmor prediction of atomic collapse was one of the strongest motivations for the Bohr model and quantum theory.
What is the dipole radiation pattern?
The power radiated per unit solid angle by an accelerating charge goes as dP/dΩ = q²a²sin²θ/(16π²ε₀c³), where θ is the angle between the acceleration vector and the direction of observation. Radiation is maximal in the plane perpendicular to the acceleration (θ = 90°) and exactly zero along the acceleration axis (θ = 0° and 180°). The pattern is a torus, or doughnut, wrapped around the acceleration direction — the same sin²θ shape as a short dipole antenna.
What is the difference between synchrotron radiation and bremsstrahlung?
Both are Larmor radiation from accelerating charges, but the acceleration has a different source. Synchrotron radiation comes from the transverse (centripetal) acceleration of charges bent by a magnetic field, as in a synchrotron storage ring or around a pulsar. Bremsstrahlung — German for 'braking radiation' — comes from the longitudinal deceleration of charges deflected by the Coulomb field of atomic nuclei, as when electrons slam into a metal target and produce X-rays. Both are computed from the relativistic (Liénard) generalization of the Larmor formula.
Is the Larmor formula valid at relativistic speeds?
No. The simple Larmor formula P = q²a²/(6πε₀c³) is the non-relativistic limit, valid only when v ≪ c. For fast particles you must use the Liénard generalization P = q²γ⁶[a² − (v×a/c)²]/(6πε₀c³), where γ = 1/√(1−v²/c²) is the Lorentz factor. This is why synchrotron radiation is so intense: at γ ≈ 10⁴ in a modern electron ring, the γ⁶ (or γ⁴ for circular motion) enhancement makes the radiated power enormous compared to the naïve estimate.
Does a charge sitting in a gravitational field radiate?
This is a famous subtlety. A charge held at rest on a table (accelerated upward by the normal force relative to a freely falling frame) versus a charge in free fall raises the question of whether the equivalence principle forces one of them to radiate. The modern consensus is that radiation is observer-dependent: a co-accelerating observer sees no radiation, while a distant inertial (free-falling) observer does. The Larmor formula is a statement about acceleration in flat spacetime; in curved spacetime the definition of 'radiation' requires care about which frame you use.