Select causal solutions with retarded Green functions
Derive retarded potentials in SI units and distinguish a causal boundary condition from the differential wave equation.
On this page
The question for today
What you’ll understand
- Normalize the three-dimensional retarded Green function for a declared wave operator.
- Use source-to-observer delay without replacing it by an instantaneous interaction.
- Explain the roles of charge conservation, gauge choice, and incoming homogeneous fields.
An antenna does not communicate a change to a distant receiver immediately. Maxwell's equations contain a finite propagation speed, but a differential equation alone does not choose whether a solution carries disturbances outward or inward. A Green function combines the local equation with that physical choice. We will keep its normalization explicit, because changing the definition of the wave operator can move factors of the speed of light between the operator, its source, and its Green function.
Fix the operator before its inverse
In vacuum, introduce scalar potential phi in volts and vector potential A in volt seconds per metre. Their derivatives give the electric and magnetic fields. The Lorenz condition couples their divergence and time derivative so that each component satisfies a sourced wave equation. This choice simplifies solving Maxwell's equations; it is not an additional physical field equation. Consistency requires charge conservation, linking the time derivative of charge density to the divergence of current density.
Use the operator with time derivatives divided by speed squared and a negative Laplacian. Its Green function has a positive spatial delta source under this convention. Integrating that function against a physical source produces the particular solution. If instead the operator were time derivative squared minus speed squared times the Laplacian, its Green function would differ by a factor of speed squared. Carrying a remembered formula between those definitions silently gives incorrect SI potentials.
Support on the past light cone
For a point impulse, the outgoing response spreads over spheres. The inverse distance factor supplies the three-dimensional geometric spreading, while the time delta restricts the response to the light cone. Its integral over source time evaluates the source at observation time minus travel time. The retarded kernel vanishes for negative elapsed time. An advanced kernel solves the same local equation with the opposite temporal support; selecting the retarded one supplies physical information about preparation and incoming waves.
For an extended source, each source position has its own distance and therefore its own retarded time. One common delay is only an approximation for a compact source viewed far away, and even then directional phase differences may matter. A static source recovers the Coulomb integral because delaying a constant changes nothing. This limit is a normalization check, not evidence that electrostatic rearrangements propagate instantly. Establishing a new static configuration still involves electromagnetic transients.
A causal particular solution and its limits
The source integral solves the inhomogeneous problem with the chosen causal prescription. One may still add a homogeneous Maxwell solution describing incident radiation or specified initial fields. Thus retarded potentials do not imply that every experiment starts with no electromagnetic field. Interfaces, conductors, and cavities require boundary conditions and often a different Green function. The free-space kernel cannot enforce a metal wall merely because the source was placed near one.
For a sinusoidal current element, a delay translates into phase. Doubling distance also reduces its vector-potential amplitude in the point-source approximation. These are separate effects: phase involves travel time, amplitude involves geometric spreading. The wave lab can isolate the addition of two equal-frequency contributions with a specified phase difference. It cannot establish causal support, create a switched pulse, or solve the three-dimensional Maxwell initial-boundary problem; those claims belong to the kernel and its assumptions.
Work it through
Convert a path difference into phase
Two compact coherent current elements drive a receiver at 1.0 MHz. Their path lengths differ by 75 m. Use c=3.0×10⁸ m/s to find the delay and phase difference, and determine the sum amplitude if their received amplitudes are equal.
The extra delay is path difference divided by speed: 75/(3.0×10⁸)=2.5×10⁻⁷ s, or one quarter of a microsecond. The farther contribution arrives later.
The period is 1.0×10⁻⁶ s, so the extra path represents one quarter cycle. Its phase lag has magnitude 90°, rather than 90 radians.
Equal received amplitudes E₀ at a quarter-cycle offset add as perpendicular phasors. Squaring the resultant gives 2E₀², so its amplitude is √2 times E₀. This assumes coherent sources and the stated equal received amplitudes.
The farther contribution lags by 0.25 μs or 90°. Their coherent sum has amplitude √2 E₀; unequal source distances do not automatically give equal received amplitudes.
The living notebook
Explore the model
Interactive lab · waves
Separate phase delay from amplitude
For each fresh run use Reset and set BOTH controls: phase 90°, amplitude ratio 1; then Reset and set phase 180°, ratio 1. Predict sum amplitudes √2 and zero before Play. The model contains two fixed-frequency, same-direction sine waves; it does not calculate path lengths or pulse causality.
- Sum at x = 0
- 0.866
- Resultant amplitude
- 1.732
- Phase difference
- 60°
Two waves share the same frequency and wave number. Playback advances t in sin(2x − t) + a sin(2x − t + phase). Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forA chosen phase delay affects coherent addition, while causal propagation requires the retarded Green function.
Pause & reason
Check your understanding
Your turn
Try it yourself
A coherent 2.0 MHz signal travels an extra 30 m in vacuum. Using c=3.0×10⁸ m/s, find its delay and phase lag. Would the same distance cause the same phase lag at 1.0 MHz?
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
First compute the time delay, then multiply its fraction of a period by 360 degrees.
Compare with the solution
The delay is 30/(3.0×10⁸)=1.0×10⁻⁷ s. At 2.0 MHz the period is 5.0×10⁻⁷ s, giving a lag of 72°. At 1.0 MHz the same geometric delay is only one tenth of a period, giving 36°. Travel time is frequency independent here, while phase lag is proportional to frequency.
Keep the idea
Take it with you
- The Green-function normalization depends on the chosen wave operator.
- Each source point is evaluated at its own retarded time.
- Retarded source fields can coexist with independently specified incident radiation.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.