All lessons

Module 4: Electromagnetic theory · Lesson 4.2 · about 50 min

Retarded and Liénard-Wiechert potentials

Radiation from moving charges for CSIR NET: retarded time, the Liénard-Wiechert velocity and acceleration fields, the Larmor and Liénard formulas, dipole antennas and radiation resistance, Thomson and Rayleigh scattering, synchrotron beaming and energy loss, and radiation reaction.

Key points, meanings and formulas on one page: choose "Save as PDF" in the print box.

Module 4 · about 50 minutes. The potentials of a moving charge, and the retarded time that goes with them, are set up in the GATE lesson on potentials and gauges. This lesson takes the next step that CSIR NET Part C keeps returning to: the fields of a moving charge, the part of them that radiates, how much power leaves, in which directions, and what that does back to the charge. Relativity enters through γ; revise it in relativistic kinematics if the factors look unfamiliar.

A charge's field at a point is set by where the charge was, how fast it moved and how it accelerated at the retarded time. The velocity part falls as 1/R² and stays attached to the charge; the acceleration part falls as 1/R and flies off as radiation. Almost every radiation formula (Larmor, Liénard, dipole antennas, Thomson scattering, synchrotron light) is that 1/R field squared and integrated over a sphere.

1. Retarded time, once more

A field point at r→ at time t hears from the charge at the earlier time tr fixed by

c(t−tr)=|r→−w→(tr)|≡𝓇

where w→(t) is the charge’s path. For any charge slower than light exactly one tr solves this: two would need the charge to keep pace with its own light. Write nˆ=𝓇→/𝓇 and β→=v→(tr)/c. The factor that appears everywhere is

κ=1−nˆ⋅β→,dtdtr=κ

The second form says what κ means: a stretch of the charge’s history dtr long reaches the observer squeezed into dt=κdtr. For a charge heading straight at you at β=0.9, κ=0.1: ten seconds of its life arrive in one. This is the same Doppler compression that turns the gentle dipole pattern of a slow charge into the narrow forward beam of a fast one.

Light spheres sent out by a moving charge. The sphere that passes through the observer now was emitted from the retarded position. Push β up and the spheres crowd ahead of the charge, so κ = 1 − n̂·β falls and the field ahead grows.

2. The Liénard-Wiechert fields

Differentiating the Liénard-Wiechert potentials (the retarded time depends on position, which is where the work lies) gives, with u→=cnˆ−v→,

E→(r→,t)=q4πϵ0𝓇(𝓇→⋅u→)3[(c2−v2)u→+𝓇→×(u→×a→)],B→=1cnˆ×E→

Everything on the right is taken at tr. Two pieces:

  • the velocity field ∝(c2−v2)u→, which falls as 1/𝓇2. For uniform motion it is the squashed Coulomb field pointing from the present position, γ times stronger sideways and γ2 times weaker ahead;
  • the acceleration field ∝𝓇→×(u→×a→), which falls as 1/𝓇 and is perpendicular to nˆ. This is the radiation.

For a slow charge (v≪c) the acceleration field reduces to

E→rad=q4πϵ0c2𝓇nˆ×(nˆ×a→),|E→rad|=qasin⁡θ4πϵ0c2𝓇

with θ measured from a→. A useful picture (J. J. Thomson’s): kink a field line by jerking the charge, and the kink travels out at c; its transverse part grows relative to the radial part in proportion to 𝓇, which is why Erad/ECoulomb=a𝓇sin⁡θ/c2.

tr
retarded time, c(t − tr) = distance from the retarded position
κ = 1 − n̂·β
compression factor dt/dtr; small when the charge moves fast towards the observer
Velocity field
the generalised Coulomb field, ∝ 1/R², carries no energy to infinity
Acceleration field
the radiation field, ∝ 1/R, transverse, E ⊥ B ⊥ n̂
Physical meaning
Energy flux S ∝ E² ∝ 1/R² from the acceleration field, times area 4πR², stays finite as R grows: energy leaves for good. That is what radiation means.

3. How much power: Larmor and Liénard

Square the slow-charge radiation field, divide by μ0c and integrate over a sphere:

dPdΩ=q2a216π2ϵ0c3sin2⁡θ,P=q2a26πϵ0c3(Larmor)

No power goes along a→; the most goes at right angles. Larmor’s formula is Lorentz invariant once written with four-vectors, and the relativistic version (Liénard) is

P=q2γ66πϵ0c3(a2−|v→×a→|2c2)

Two special cases do most exam work:

Motion Force-acceleration link Power
linear (a→∥v→) F=γ3ma P=q2γ6a26πϵ0c3=q2F26πϵ0m2c3
circular (a→⟂v→) F=γma P=q2γ4a26πϵ0c3=q2γ2F26πϵ0m2c3

For the same force a circular orbit radiates γ2 times more. And everything scales as 1/m2: protons radiate (1836)2≈3.4×106 times less than electrons under the same force, which is why the largest proton machines are circular and the largest electron machines will be linear.

Worked example: is a linac wasteful?

Set-up. Electrons gain 10 MeV per metre in a linear accelerator, so F = 10 MeV/m = 1.60 × 10⁻¹² N, at any γ.

Power radiated. P = e²F²/(6πε₀m²c³) = (2/3) re F²/(mc), with re = e²/4πε₀mc² = 2.818 fm.

Compare with the energy gain per second, Fc: P/(Fc) = (2/3) re F/(mc²) = (2/3) × 2.818 × 10⁻¹⁵ m × (10 MeV/m)/(0.511 MeV) = 3.7 × 10⁻¹⁴.

Answer. Radiation loss in a linac is utterly negligible; it would matter only for fields near mc²/re ≈ 2 × 10²⁰ V/m.

Synchrotron radiation

For a ring of radius R, a=v2/R, and the energy lost per turn is P⋅2πR/v:

ΔEturn=q2β3γ43ϵ0R,ΔEturn[keV]≈88.5E4[GeV]R[m]  (electrons)

A 3 GeV electron ring with a 10 m bending radius loses about 717 keV a turn, supplied by RF cavities. The light comes out in a forward cone of half-angle about 1/γ (the κ−5 factor in the relativistic angular distribution), so an observer sees a flash lasting about R/(cγ3) and a spectrum reaching up to the critical frequency ωc=32γ3c/R: from a 3 GeV ring with R = 10 m, ħωc ≈ 6 keV, X-rays.

4. Dipoles and antennas

An oscillating dipole p(t)=p0cos⁡ωt is a charge q with a=−ω2x. Larmor’s formula averaged over a cycle gives

⟨P⟩=μ0p02ω412πc,d⟨P⟩dΩ=μ0p02ω432π2csin2⁡θ

The ω4 is the heart of the matter. A short centre-fed antenna of length l≪λ carrying a uniform current I0cos⁡ωt has p0ω=I0l, and defining ⟨P⟩=12I02Rrad,

Rrad=2π3Z0(lλ)2≈789(lλ)2 Ω

with Z0=μ0c=376.7 Ω. A 1 m whip at 15 MHz (λ=20 m) has Rrad≈2 Ω, so most of the power fed to it is lost as heat in the wire and the matching coil: short antennas are inefficient. (If the current tapers to zero at the ends, as in a real short dipole, Rrad drops by a factor 4.) A half-wave dipole is not short, and its exact resistance is 73 Ω.

The magnetic dipole m0cos⁡ωt radiates ⟨P⟩=μ0m02ω4/12πc3, with E→ and B→ swapped relative to the electric case. For a loop of size b the ratio to an electric dipole of similar size is (ωb/c)2: small loops are weak radiators.

5. Scattering of light by charges

Free electron (Thomson). A wave of amplitude E0 shakes an electron with a=eE0/m. Radiated power over incident intensity is

σT=8π3re2=6.652×10−29 m2=0.665 b

independent of frequency. Unpolarised light scatters with angular weight 12(1+cos2⁡θ), and at 90° the scattered light is completely polarised.

Bound electron (Lorentz model). An electron on a spring of frequency ω0 responds with amplitude ∝1/(ω02−ω2), so

σ(ω)=σTω4(ω02−ω2)2+γd2ω2

Far below resonance (ω≪ω0) this is Rayleigh scattering, σ≈σT(ω/ω0)4: violet light at 400 nm scatters (700/400)4=9.4 times more than red at 700 nm, and the sky is blue. Far above, the electron behaves as free and σ→σT.

6. Radiation reaction

A radiating charge must lose energy, so something pushes back. Balancing the Larmor power on average gives the Abraham-Lorentz force

F→rad=mτa→˙,τ=q26πϵ0mc3=2re3c=6.26×10−24 s (electron)

τ is the light-crossing time of a classical electron, so radiation reaction matters only on time scales that short or for motion sustained over many cycles. Two consequences:

  • Line width. A classical oscillator loses energy at the rate Γ=ω02τ, so its line has width Δλ=2πcτ=4π3re=1.18×10−14 m, the same at every wavelength.
  • The classical atom collapses. An electron spiralling in from the Bohr radius radiates all the way and reaches the nucleus in t=a03/(4re2c)≈1.6×10−11 s. Classical physics cannot make a stable atom; quantum mechanics does.

The Abraham-Lorentz equation has runaway solutions (a∝et/τ) and, if they are excluded, pre-acceleration that starts a time τ before the force. These signal that a point charge is not a consistent classical object; they never matter at measurable scales.

“Any moving charge radiates.” Only an accelerated charge radiates. A charge in uniform motion carries its field along: in its rest frame there is nothing but a Coulomb field, and a Lorentz boost cannot create an energy flux that escapes to infinity. The current in a straight DC wire does not radiate either. What does radiate is any change of velocity, including a change of direction at constant speed: a circular orbit radiates at all times.

  • Retarded time: c(t−tr)=𝓇; compression factor κ=1−nˆ⋅β→=dt/dtr.
  • LW fields: velocity part ∝1/𝓇2, acceleration part ∝1/𝓇 and transverse; B→=nˆ×E→/c.
  • Larmor P=q2a2/6πϵ0c3, pattern sin2⁡θ. Liénard P∝γ6(a2−|v→×a→|2/c2): linear γ6a2, circular γ4a2; for equal force circular wins by γ2; everything ∝1/m2.
  • Synchrotron: ΔEturn=q2β3γ4/3ϵ0R (88.5 keV E4[GeV]/R[m] for electrons), cone 1/γ, ωc=32γ3c/R.
  • Dipole ⟨P⟩=μ0p02ω4/12πc; short antenna Rrad=789(l/λ)2 Ω; half-wave dipole 73 Ω.
  • Thomson σT=8π3re2=0.665 b, frequency independent; Rayleigh σ∝ω4 below resonance.
  • Radiation reaction F=mτa˙, τ=6.26×10−24 s; classical line width 1.18 × 10⁻¹⁴ m; classical atom lifetime 1.6 × 10⁻¹¹ s.

Practise: exam questions on this lesson

NET Part C style, each with a full solution. More on the CSIR NET Physics practice page.

Next: plasma oscillations and dispersion, where the same shaken electrons act together and turn a medium opaque below a cutoff frequency. For waves in ordinary media see Maxwell’s equations and waves.

Check yourself

1An electron in a linear accelerator and one in a circular ring feel the same force F perpendicular or parallel to v. At large γ, which radiates more?
2How does the Thomson cross-section of a free electron depend on the frequency of the incident light (well below mc²/h)?
3The far-field radiation from a charge depends on which property of its motion at the retarded time?

Sources

  • D. J. Griffiths, Introduction to Electrodynamics, 4th ed. (2017), sections 10.3 and 11.1 to 11.2
  • J. D. Jackson, Classical Electrodynamics, 3rd ed. (1999), chapters 9, 14 and 16
  • A. Zangwill, Modern Electrodynamics (2013), chapters 20 and 23
  • CSIR-UGC NET Physical Sciences syllabus, Electromagnetic Theory unit, published by the National Testing Agency

Your ticks stay in this browser. Save them as a file on the course home.

Updated 2026-11-22. Not affiliated with CSIR, UGC or the National Testing Agency. Check the official CSIR NET website for the current pattern. All questions are original.