Liénard–Wiechert Potentials

The retarded potentials answer any source; now feed them the most important source in physics — a single point charge on an arbitrary trajectory $\mathbf{w}(t)$ — and watch the machinery produce a surprise. The tempting answer is obvious: a point charge should give $V = q/4\pi\varepsilon_0\mathscr{r}$ with $\mathscr{r}$ read off at the retarded time. The tempting answer is wrong, and wrong by a factor that can reach a hundred million. The correct result, found by Alfred-Marie Liénard in 1898 and Emil Wiechert in 1900, carries an extra denominator:

V = \frac{1}{4\pi\varepsilon_0}\,\frac{qc}{\mathscr{r}c - \boldsymbol{\mathscr{r}}\cdot\mathbf{v}}, \qquad \mathbf{A} = \frac{\mathbf{v}}{c^2}\,V,

everything evaluated at the retarded time. The factor $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}/c)^{-1}$ is pure retardation geometry — no Lorentz contraction, no time dilation, no relativity at all — yet it is the seed of every relativistic beaming phenomenon in electrodynamics: the headlight of synchrotron light sources, the forward spike of bremsstrahlung, the blinding sweep of pulsar beams. Where does it come from? From counting. The retarded integral collects the charge's announcements over the observer's past light cone, and a charge moving toward you crowds more of its history into your collection window — like a train that stays longer in a photograph the faster it approaches the camera whose shutter sweeps toward it. This article earns the factor twice — once by honest geometry, once by honest delta-function calculus — proves there is only ever one retarded moment to collect (for $v < c$ ; the failure of that theorem in dielectrics is Cherenkov radiation), and packages the result for the two articles that will differentiate it into the complete fields of a moving charge.

Key takeaways

For a point charge $q$ on trajectory $\mathbf{w}(t)$ , the retarded integrals collapse to the Liénard–Wiechert potentials: V(\mathbf{r},t) = \frac{1}{4\pi\varepsilon_0}\frac{qc}{\mathscr{r}c - \boldsymbol{\mathscr{r}}\cdot\mathbf{v}}, \qquad \mathbf{A}(\mathbf{r},t) = \frac{\mathbf{v}}{c^2}\,V = \frac{\mu_0}{4\pi}\frac{qc\,\mathbf{v}}{\mathscr{r}c - \boldsymbol{\mathscr{r}}\cdot\mathbf{v}}, with $\boldsymbol{\mathscr{r}} = \mathbf{r} - \mathbf{w}(t_r)$ and $\mathbf{v} = \dot{\mathbf{w}}(t_r)$ both at the retarded time.
The naive $q/4\pi\varepsilon_0\mathscr{r}$ is wrong: evaluating a delta function whose argument moves with $t_r$ produces the Jacobian $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})^{-1}$ — an apparent-volume effect, present even for constant velocity.
The retarded time is unique: for $v < c$ the past light cone of any event intersects the worldline exactly once — a charge cannot outrun its own announcements. (In a medium where $v > c/n$ it can: two retarded times, a Mach cone, Cherenkov light.)
The factor is not relativistic in origin — no $\gamma$ appears; it is retardation bookkeeping, though it becomes the engine of relativistic beaming: head-on at $\beta = 0.99$ the enhancement is 100; for a 10 GeV electron, $2\gamma^2 \approx 7.7\times10^8$ .
The enhanced potential does not mean enhanced charge: Gauss's law still reports exactly $q$ — the factor redistributes the potential's landscape, it does not mint charge.
The pair packages covariantly as $A^\mu = \tfrac{\mu_0 qc}{4\pi}\,u^\mu/(R_\nu u^\nu)$ — a preview cashed in the relativity chapter and the Jackson-register treatment.
Key idea: the point-charge retarded integral collapses with a Jacobian: the potentials of a moving charge are the static ones divided by $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}/c)$ , evaluated at the unique retarded time — geometry's own beaming factor.

The core insight

The retarded integral for a point charge collapses with the Jacobian factor $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}/c)^{-1}$ : the Liénard–Wiechert potentials are the static potentials enhanced by pure retardation geometry, evaluated at the unique retarded time.

Where this lives today

The LW denominator is a multi-billion-euro industry. Synchrotron light sources — the ESRF, APS, and their fourth-generation successors — and X-ray free-electron lasers like the European XFEL and LCLS are machines for pointing the $(1-\hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})^{-1}$ factor at a sample: multi-GeV electrons steered through undulators, their retarded potentials' forward pileup delivering X-ray brilliance $10^{10}$ times anything a laboratory tube manages, illuminating everything from protein crystallography to operando battery chemistry. Accelerator codes evaluate LW potentials directly for coherent synchrotron radiation and beam–beam wakefields, where the "who saw whom at which retarded time" bookkeeping of this article is the computational bottleneck. Astrophysics runs the same formula at nature's scale. Pulsar radio beams sweeping the sky, relativistic jets from active galactic nuclei whose approaching sides outshine their receding ones a thousandfold (Doppler favoritism with the same denominator), and the curvature radiation of magnetospheric plasmas all wear LW clothing. Even the uniqueness theorem's loophole is industrial: Cherenkov detectors — the photomultiplied water tanks of IceCube and Super-Kamiokande — watch for the double-retarded-time Mach cones of particles beating light's medium speed, catching neutrinos by the failure mode of this article's central theorem.

Origins · 1 minAlfred-Marie Liénard and how this was discovered

Alfred-Marie Liénard

1869–1958 · École des Mines, Paris — who solved the moving point charge seven years before relativity

Liénard was a French mining engineer and physicist — later director of the École des Mines de Paris — who in 1898 published, in the journal L'Éclairage Électrique ("Electric Lighting"; the venue itself a period piece), the exact retarded potentials of an arbitrarily moving point charge, complete with the velocity denominator this article is named for, and pushed on to the radiated power of an accelerated charge including the relativistic velocity factors (the "Liénard result" the chapter meets again with the relativistic Larmor formula). All of it pre-relativity: light-cone geometry plus Lorenz's retarded potentials, no $\gamma$ , no Einstein — seven years early. Emil Wiechert, the Göttingen geophysicist (a founder of modern seismology and namesake of the Wiechert seismograph, still displayed running in Göttingen), derived the same potentials independently in 1900 by the delta-function-style route. The double name is deserved: two communities, two methods, one factor. Liénard spent his later career on engineering mathematics — the Liénard equation of nonlinear oscillator theory is also his — a reminder that the deepest formula of radiation theory came from a man whose day job was mining machinery.

History

1867

— Lorenz's retarded potentials: the master integrals await a worthy source

1898

— Liénard evaluates them for an arbitrarily moving point charge; publishes the beaming denominator and the relativistic power formula

1900

— Wiechert, independently at Göttingen, derives the same potentials via the delta-collapse route

1904

— Schwarzschild's covariant-style formulation; the potentials enter the relativity conversation Einstein is about to start

1912

— Schott's treatise makes the LW potentials the engine of classical radiation theory

1934

— Cherenkov observes the blue glow of the uniqueness theorem's loophole; Frank and Tamm explain it in 1937

1947

— General Electric technicians see synchrotron light — the LW forward pileup — by eye in an accelerator hall

2009

— LCLS lases: the X-ray free-electron laser, the beaming denominator's most extreme industrialization

The physical picture

Start with the theorem that makes "the retarded time" a well-defined phrase. Sit at the observation event $(\mathbf{r}, t)$ and shrink your past light cone backward: a sphere of radius $c(t - t')$ collapsing at the speed of light onto your location. The charge's worldline must cross this collapsing sphere — the sphere starts infinitely large and ends at a point — and it can cross only once: to be caught twice, the charge would have to duck inside the shrinking sphere and then get back outside it, which means outrunning a surface that moves at $c$ . For $v < c$ , impossible. One crossing, one retarded moment $t_r$ , one term in the potential. (Hold that argument; its failure in matter, where the field's speed is $c/n$ and a charge may legally exceed it, is priced in the FAQ — the double-crossing is Cherenkov radiation.) Now the factor. The retarded integral does not sample the charge at an instant — it sweeps the collapsing sphere through the charge, collecting whatever charge density the sphere encounters during its passage. For an extended blob at rest, the sweep crosses each bit exactly once and total collected charge is $q$ : the naive answer. But let the blob move toward you while the sphere sweeps inward through it: blob and sphere approach each other's territory in opposite senses — the sphere spends longer inside the blob than the blob's rest length suggests, collecting announcements from a stretch of the blob's history of length $d/(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})$ rather than $d$ . The collection is over-complete for approach, under-complete for recession. Shrink the blob to a point and the over-collection survives as a finite factor: the apparent-volume enhancement. Nothing moved faster than light; nothing Lorentz-contracted; the counting changed.

Figure

The information-collecting sphere — a light-speed sphere collapsing onto the observer sweeps through the charge; motion toward the observer stretches the collection

The potentials

The objects, with the notation the rest of the chapter will lean on. The charge's trajectory is $\mathbf{w}(t)$ ; the retarded time $t_r(\mathbf{r},t)$ solves the light-cone condition; the retarded separation runs from the retarded position to the field point:

|\mathbf{r} - \mathbf{w}(t_r)| = c\,(t - t_r), \qquad \boldsymbol{\mathscr{r}} \equiv \mathbf{r} - \mathbf{w}(t_r), \qquad \mathbf{v} \equiv \dot{\mathbf{w}}(t_r), \qquad \boldsymbol{\beta} \equiv \mathbf{v}/c. \qquad (1)

where: the implicit equation for $t_r$ has exactly one solution for $v < c$ (the collapsing-sphere argument above, made algebraic: the function $f(t') = |\mathbf{r}-\mathbf{w}(t')| - c(t-t')$ is strictly monotonic, since $|d|\mathbf{r}-\mathbf{w}|/dt'| \le v < c$ while the second term changes at rate $c$ ). Everything that follows is evaluated at this one moment. The Liénard–Wiechert potentials:

V(\mathbf{r},t) = \frac{1}{4\pi\varepsilon_0}\,\frac{qc}{\mathscr{r}c - \boldsymbol{\mathscr{r}}\cdot\mathbf{v}} = \frac{1}{4\pi\varepsilon_0}\,\frac{q}{\mathscr{r}\,(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})}, \qquad \mathbf{A}(\mathbf{r},t) = \frac{\mathbf{v}}{c^2}\,V(\mathbf{r},t). \qquad (2)

where: the denominator correction $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})$ is the article's protagonist: unity for a static charge (Coulomb recovered), less than unity for approach (potential enhanced), greater for recession (suppressed), and never zero in vacuum since $|\hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta}| \le \beta < 1$ . Note the tight kinship of the pair: $\mathbf{A} = \mathbf{v}V/c^2$ — the vector potential is the scalar potential riding the charge's velocity, one structure, four components, begging for the four-vector packaging delivered below. The two derivations now earn (2) by independent routes; if they agreed by accident, the agreement would be the most elaborate coincidence in the chapter.

Derivation AThe apparent-volume route — geometry earns the factor

The goal is the factor $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})^{-1}$ from pure retardation geometry, by modeling the point charge honestly as the limit of an extended blob and counting what the retarded integral actually collects. The strategy is Griffiths's train argument made quantitative: fix one observation event, follow the collection sphere through the blob, and compare the swept extent of the blob to its geometric extent. No relativity enters — lengths are lengths, one frame throughout. This is the physical derivation; the delta-function calculus below is its formal cross-check.

Step 1. Set the sweep. Let the blob have thickness $d$ along the line of sight $\hat{\boldsymbol{\mathscr{r}}}$ (pointing from charge to observer) and move with velocity $\mathbf{v}$ . Announcements arriving at the observer simultaneously left the far edge earlier than the near edge — earlier by the light-travel difference across the region the sphere sweeps. Call the swept length $L$ : the far-edge signal departs $\Delta t = L/c$ before the near-edge signal.

Step 2. The blob moves during the sweep. In the interval $\Delta t$ , the blob advances $v\,\Delta t\cos\theta = (L/c)\,\hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}$ along the line of sight, where $\theta$ is the angle between $\mathbf{v}$ and $\hat{\boldsymbol{\mathscr{r}}}$ . The sweep must cover the blob's geometric thickness plus the ground the blob gained toward the sphere:

L = d + \frac{L}{c}\,\hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v} \qquad\Longrightarrow\qquad L = \frac{d}{1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta}}.

Why the sign runs the intuition: approaching blob ( $\hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta} > 0$ ) chases the collapsing sphere, so the sphere needs extra travel to finish the crossing — longer sweep, more history collected. A receding blob backs away from the sweep and is crossed quickly. The train analogy: a train driving toward a camera whose focal "shutter surface" sweeps toward it stays in frame longer than its rest length warrants.

Step 3. Count the collection. The retarded integral gathers $1/\mathscr{r}$ -weighted charge over the swept region: effectively the blob's charge multiplied by $L/d$ , since every slab of the blob is crossed once but the history collected corresponds to the stretched extent. In the point limit ( $d \to 0$ at fixed $q$ ), $\mathscr{r}$ is common to all contributions and

V = \frac{1}{4\pi\varepsilon_0}\,\frac{q}{\mathscr{r}}\cdot\frac{1}{1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta}}.

Why this is not charge non-conservation: the observer's sphere collects announcements — samples of the charge's history — not the charge itself. An over-complete collection of a conserved charge's past is perfectly legal; Gauss's law, applied to the resulting fields at any instant, still encloses exactly $q$ (worked example 1 checks a case). The factor is epistemic geometry: how much worldline fits on one light cone.

Step 4. The vector potential rides along. The current density of the point charge is $\mathbf{J} = \rho\mathbf{v}$ , so the identical collection with an extra factor $\mathbf{v}(t_r)/c^2$ (constant over the collapsed integral) gives $\mathbf{A} = \mathbf{v}V/c^2$ — equation (2) complete. Verified: matches Griffiths §10.3.1's train argument and Eq. 10.46; static limit $\boldsymbol{\beta}=0$ recovers Coulomb; the factor depends only on the velocity — not acceleration — at $t_r$ , as the geometry demands.

Go deeper · Derivation B · 1 minThe delta-function route — calculus cross-checks the geometry

Derivation BThe delta-function route — calculus cross-checks the geometry

The goal is the same factor from the formal side: insert the point-charge densities $\rho = q\,\delta^3(\mathbf{r}' - \mathbf{w}(t_r))$ , $\mathbf{J} = \rho\,\dot{\mathbf{w}}(t_r)$ into the retarded integrals and evaluate the delta exactly. The subtlety — the crux of the whole computation — is that the delta's argument contains the integration variable twice: explicitly in $\mathbf{r}'$ , and implicitly through $t_r = t - |\mathbf{r}-\mathbf{r}'|/c$ . A composed delta pays a Jacobian, and the Jacobian is the theorem. The clean route goes through an auxiliary time integral, which linearizes the composition.

Step 1. Un-compose with an auxiliary delta. Write the retarded density as a time integral against the pinning delta (the Green-function structure of the chapter, used in reverse):

V = \frac{1}{4\pi\varepsilon_0}\int\!\!\int \frac{q\,\delta^3(\mathbf{r}' - \mathbf{w}(t'))}{|\mathbf{r}-\mathbf{r}'|}\,\delta\!\left(t' - t + \frac{|\mathbf{r}-\mathbf{r}'|}{c}\right)dt'\,d^3r'.

Why this helps: at fixed $t'$ , the spatial delta is now innocent — its argument is linear in $\mathbf{r}'$ — so the $d^3r'$ integral collapses without penalty, placing $\mathbf{r}' = \mathbf{w}(t')$ everywhere.

Step 2. Collapse space, confront time:

V = \frac{q}{4\pi\varepsilon_0}\int \frac{\delta\!\left(t' - t + |\mathbf{r}-\mathbf{w}(t')|/c\right)}{|\mathbf{r}-\mathbf{w}(t')|}\,dt' .

The remaining delta is genuinely composed: its argument $g(t') = t' - t + |\mathbf{r}-\mathbf{w}(t')|/c$ is a nonlinear function of $t'$ .

Step 3. Pay the Jacobian. The composition rule $\delta(g(t')) = \delta(t' - t_r)/|g'(t_r)|$ requires the root (unique — the monotonicity argument of equation (1); indeed $g' > 0$ always, so the absolute value is redundant in vacuum) and the derivative:

g'(t') = 1 + \frac{1}{c}\frac{d}{dt'}|\mathbf{r}-\mathbf{w}(t')| = 1 - \frac{\hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}}{c},

since the distance to a point retreating along $-\mathbf{v}$ changes at rate $-\hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}$ . Why this is the same physics as Derivation A: $g'$ measures how fast the light-cone condition scans across the worldline — the sweep rate of the collection sphere through the charge's history. Slow scan (approach) means the delta spends "more time" near its root per unit observer time: the identical stretching, now spoken in calculus.

Step 4. Assemble. The $t'$ integral collapses onto $t_r$ with the factor $1/g'(t_r)$ :

V = \frac{1}{4\pi\varepsilon_0}\,\frac{q}{\mathscr{r}\left(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta}\right)}, \qquad \mathbf{A} = \frac{\mu_0}{4\pi}\,\frac{qc\,\boldsymbol{\beta}}{\mathscr{r}\left(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta}\right)} = \frac{\mathbf{v}}{c^2}V,

with every symbol at $t_r$ . Two routes, one factor: the geometry is the calculus, the calculus is the geometry. Verified: matches Griffiths Eq. 10.46–10.47 and Jackson Eq. 14.8; the same auxiliary-delta manipulation is the standard opening of every covariant treatment.

The covariant package

The pair (2) compresses into one relativistic line — stated here as a preview, since the four-vector machinery belongs to the relativity chapter and the fully covariant working of this subject to the Jackson-register treatment at this track's end. With the four-potential $A^\mu = (V/c, \mathbf{A})$ , the charge's four-velocity $u^\mu = \gamma(c, \mathbf{v})$ , and the null separation $R^\mu = (\mathscr{r}, \boldsymbol{\mathscr{r}})$ from the retarded event to the field event ( $R^\mu R_\mu = \mathscr{r}^2 - |\boldsymbol{\mathscr{r}}|^2 = 0$ — the light-cone condition wearing index notation):

A^\mu(x) = \frac{\mu_0\,qc}{4\pi}\,\frac{u^\mu}{R_\nu u^\nu}\bigg|_{\text{ret}}, \qquad R_\nu u^\nu = \gamma\left(\mathscr{r}c - \boldsymbol{\mathscr{r}}\cdot\mathbf{v}\right). \qquad (3)

where: the $\gamma$ 's cancel between numerator and denominator when components are extracted — check the time component: $A^0 = \tfrac{\mu_0 qc}{4\pi}\,\gamma c/[\gamma(\mathscr{r}c - \boldsymbol{\mathscr{r}}\cdot\mathbf{v})] = V/c$ ✓. Two lessons ride on (3). First, the form is manifestly covariant: potentials of a moving charge in any frame, one formula. Second — the sharper lesson — no $\gamma$ survives in (2): the Liénard–Wiechert factor is not a Lorentz-transformation effect but light-cone geometry, present in a strictly pre-relativistic reading of retardation (Liénard had it seven years before Einstein). Relativity organizes the result; retardation produces it.

Limiting cases

case	result	meaning
charge at rest	$V = q/4\pi\varepsilon_0\mathscr{r}$ , $\mathbf{A} = 0$	Coulomb recovered; the factor is unity
slow motion ( $\beta \ll 1$ )	$V \approx \tfrac{q}{4\pi\varepsilon_0\mathscr{r}}(1 + \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})$ , $\mathbf{A} \approx \tfrac{\mu_0 q\mathbf{v}}{4\pi\mathscr{r}}$	first-order beaming; $\mathbf{A}$ is the elementary current's potential
constant velocity	closed form in present position: $V = \tfrac{q}{4\pi\varepsilon_0 R}\left(1 - \beta^2\sin^2\theta\right)^{-1/2}$	the retarded bookkeeping conspires to point at the charge now — worked example 1
head-on approach	$V$ enhanced by $1/(1-\beta)$ → $2\gamma^2$ ultrarelativistically	the beaming ledger — worked example 2
$v \to c/n$ in a medium	denominator can vanish; two retarded times beyond	the Cherenkov threshold: the uniqueness theorem's legal loophole (FAQ)

Common trapThree traps around one denominator

Trap 1 — the missing factor: writing $V = q/4\pi\varepsilon_0\mathscr{r}_{\text{ret}}$ . The delta collapse pays a Jacobian; forget it and every fast-charge result that follows (fields, radiation, beaming) is wrong by up to orders of magnitude. Trap 2 — the wrong attribution: calling the factor "Lorentz contraction" or "a Doppler illusion." No $\gamma$ appears; the effect is apparent-volume counting on the light cone, present for a uniformly gliding charge that radiates nothing and contracts nothing in the observer's bookkeeping of potentials. (It is related to the Doppler factor — same denominator structure — but it multiplies amplitudes at one event, not frequencies between events.) Trap 3 — the over-reading: concluding that an approaching charge "carries more charge." Charge is invariant and Gauss's law over any surrounding surface still yields exactly $q$ ; the factor reshapes the potential landscape — crowding equipotentials forward — without minting a single coulomb. When a formula tempts you to violate charge invariance, the error is always in the reading, never in the ledger.

Worked examples

Figure

Constant velocity, beta = 0.8 — the retarded machinery centers the potential on the PRESENT position, with equipotentials bunched broadside by 1.667 and exactly Coulomb fore-and-aft

1 · Constant velocity — the potential that points at the present (confidence builder)

Setup. A charge moves with constant velocity $\mathbf{v} = v\hat{\mathbf{x}}$ , passing the origin at $t = 0$ . Evaluate the LW scalar potential and express it in terms of the present position — Griffiths Ex. 10.4 — then put numbers at $\beta = 0.8$ .

Step 1. The retarded bookkeeping. With $\mathbf{w}(t) = vt\,\hat{\mathbf{x}}$ , the light-cone condition gives a quadratic for $t_r$ ; selecting the past root and grinding the algebra of (2) yields the famous closed form

V(\mathbf{r},t) = \frac{1}{4\pi\varepsilon_0}\,\frac{q}{R\,\sqrt{1 - \beta^2\sin^2\theta}},

where $R$ and $\theta$ are the distance and polar angle measured from the charge's present position $vt\,\hat{\mathbf{x}}$ .

Step 2. Marvel correctly. The potential of a uniformly moving charge is centered on where the charge is, not where it was — the retarded machinery, fed a predictable trajectory, extrapolates perfectly. No paradox: the light cone samples the past, but for constant velocity the past determines the present, and the algebra performs the extrapolation automatically. (This present-pointing survives into the fields — next article — and its rude interruption when the charge accelerates is exactly where radiation comes from.)

Step 3. Numbers at $\beta = 0.8$ . Broadside ( $\theta = 90°$ ): $1/\sqrt{1 - 0.64} = \mathbf{1.667}$ — the potential is two-thirds stronger than Coulomb at the same present distance. Fore-and-aft ( $\theta = 0$ or $180°$ ): factor $1/\sqrt{1-0} = \mathbf{1.000}$ — exactly Coulomb. The landscape is a pancake: equipotentials bunch toward the transverse plane, the germ of the flattened field pattern the next article will draw.

Step 4. The Gauss audit. Integrate the resulting $\mathbf{E}$ (next article's Eq. for constant velocity) over a sphere around the present position: the enhancement at the equator and the deficit at the poles average to exactly $q/\varepsilon_0$ of flux. Charge invariant, books closed.

The lesson: for uniform motion the LW potentials are Coulomb-in-disguise — anisotropically dressed, present-centered, and flux-honest. All genuine news in electrodynamics is news of acceleration.

2 · The beaming ledger — from lab bench to light source (exam level)

Setup. Tabulate the LW enhancement $D^{-1} = (1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})^{-1}$ for a charge headed straight at the observer, versus straight away, and push to accelerator energies.

Step 1. The modest regime. $\beta = 0.5$ : forward $1/(1-0.5) = 2.0$ , backward $1/(1.5) = 0.667$ — a 3:1 fore-aft asymmetry. $\beta = 0.9$ : forward $\mathbf{10}$ , backward $0.526$ — 19:1. $\beta = 0.99$ : forward $\mathbf{100}$ , backward $0.502$ — 199:1. The denominator is a precision instrument for converting speed into anisotropy.

Step 2. The accelerator regime. A 10 GeV electron (a synchrotron light source's storage ring): $\gamma = E/mc^2 = 10^{10}/(5.11\times10^5) = 1.96\times10^4$ . Ultrarelativistically $1 - \beta \approx 1/2\gamma^2$ , so the head-on enhancement is

\frac{1}{1-\beta} \approx 2\gamma^2 = 2\,(1.96\times10^4)^2 = \mathbf{7.7\times10^8}.

Nearly a billion-fold amplification of the forward potential — from bookkeeping the naive $q/4\pi\varepsilon_0\mathscr{r}$ misses entirely.

Step 3. Where the ledger cashes out. The fields inherit powers of this denominator ( $D^{-3}$ for the field of the next article), and the radiated power angular distribution inherits up to $D^{-5}$ (the relativistic Larmor story later in the chapter): the forward spike narrows to an angle $\sim 1/\gamma \approx 50\ \mu$ rad here. That collimation is the product: a synchrotron beamline is the LW denominator, engineered — an X-ray searchlight made of pure retardation geometry.

The lesson: every factor of $(1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta})$ in the coming articles traces back to this one Jacobian. Learn it here, spend it everywhere.

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., §10.3.1 · Jackson, Classical Electrodynamics, 3rd ed., §14.1 · Zangwill, Modern Electrodynamics, §23.2 · Purcell & Morin, Electricity and Magnetism, 3rd ed., Ch. 5 (the constant-velocity field) · Liénard, L'Éclairage Électrique 16 (1898); Wiechert, Ann. Phys. 309 (1901)

Provenance. Equations follow Griffiths §10.3.1 (Eqs. 10.46–10.47) and Jackson §14.1 (Eq. 14.8). A verifier should: (1) reproduce the delta-function collapse including the Jacobian $g'(t_r) = 1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\boldsymbol{\beta}$ ; (2) confirm the geometric apparent-volume argument yields the same factor; (3) check the static and slow-motion limits; (4) verify the covariant form's time component recovers $V/c$ with the $\gamma$ 's cancelling; (5) confirm the constant-velocity closed form (Griffiths Ex. 10.4). Addendum for this edition: the uniqueness-of- $t_r$ theorem is stated with its monotonicity proof in equation (1)'s commentary and spent in Derivation B (Step 3); the covariant compactification (3) is held at preview level — the four-potential machinery belongs to the relativity chapter (The Four-Potential) and the fully covariant working to the Jackson-register twin at this track's end (within-track title dupe, accepted per the track recon; that article owns the covariant field tensor route). Differentiation into fields is wholly deferred to Fields of a Moving Point Charge. Worked-example numbers ( $1.667$ at $\beta=0.8$ broadside; forward/backward ladder $2/10/100$ vs $0.667/0.526/0.502$ ; $\gamma = 1.96\times10^4$ , $2\gamma^2 = 7.7\times10^8$ for 10 GeV) recomputed for this edition. ⚠ Slug note: bare `lienard-wiechert-potentials` collides with an old physics-wiki row — REPLACE at global cutover per recon §c-A; the aq- prefix protects until then.

Builds on: Retarded Potentials · Lorenz Gauge · Inhomogeneous Wave Equations for Potentials · Jefimenko's Equations Leads to: Fields of a Moving Point Charge · Electric Dipole Radiation · Relativistic Larmor Formula (Liénard's Result) · Liénard–Wiechert Potentials (covariant treatment, E11)

Is the Liénard-Wiechert factor just the Doppler effect in disguise?

Sibling, not disguise. The Doppler factor $(1 - \hat{\mathbf{n}}\cdot\boldsymbol{\beta})^{-1}$ compares rates between source proper events and reception events — how many wave crests per observer second. The LW factor has the identical algebraic form because it comes from the same light-cone geometry (the $g'(t_r)$ of Derivation B is exactly the crest-compression rate), but it does a different job: it multiplies the amplitude of the potential at a single reception event, converting collected worldline history into field strength. The distinction cashes out in exponents: frequency shifts carry one power of the factor, LW fields carry three, radiated power patterns up to five. Conflating them predicts wrong powers of $(1-\beta)$ in every beaming problem — a classic exam ambush. Common cause, different observables.

If the potential is enhanced by a factor of a hundred million, why doesn't Gauss's law see extra charge?

Because the enhancement never touches the flux integral's bookkeeping. Charge is a Lorentz invariant and a conserved scalar; the LW factor redistributes the potential (and hence field) pattern — piling it into the forward direction, thinning it aft — while the solid-angle-weighted total over any closed surface remains pinned to $q/\varepsilon_0$ . Worked example 1's constant-velocity case shows the mechanism: broadside enhancement $\times1.667$ at $\beta = 0.8$ against fore–aft factor $1$ , and the anisotropic field integrates to exactly the Coulomb flux. The general moral: potentials and fields are samples of the charge's light-cone history and may be arbitrarily lopsided; charge itself is an integral over a spacelike surface and is untouchable. Any argument that converts beaming into charge creation has confused a sampling density with a conserved quantity.

Where exactly did the derivation use v less than c, and what breaks at the Cherenkov threshold?

Twice, at the same joint. The uniqueness of $t_r$ needed the light-cone function $g(t')$ strictly monotonic, i.e. $|d|\mathbf{r}-\mathbf{w}|/dt'| < c$ ; and Derivation B's Jacobian needed $g'(t_r) = 1 - \hat{\boldsymbol{\mathscr{r}}}\cdot\mathbf{v}/c > 0$ . In vacuum both hold for any massive particle. In a dielectric the fields propagate at $c/n$ , so the effective condition is $v < c/n$ — and a fast particle in water ( $n = 1.33$ ) can violate it. Beyond threshold, $g$ is non-monotonic: two retarded times contribute (the observer hears the approach and the recession of the "sonic boom" simultaneously), and on the Mach cone itself $g' = 0$ — the potential formally diverges as the collected history piles onto a single arrival. That singular pileup, smoothed by the medium's dispersion, is Cherenkov radiation: blue light in reactor pools, the trigger signal of neutrino telescopes. The failure mode of this article's tidiest theorem is a working detector technology.

Does the constant-velocity charge's present-centered potential mean the field responds to the present instantaneously?

No — it means prediction is not communication. The retarded machinery samples only the past light cone; for uniform motion, that past implies the present position, and the algebra of worked example 1 performs the implication. The distinction becomes physical the instant the implication fails: let the charge swerve at time $t_0$ , and observers beyond $c(t-t_0)$ continue to see fields pointing at the extrapolated position — where the charge would have been — not the actual one. The disagreement between extrapolation and actuality propagates outward at $c$ as a kink in the field lines, and that kink is precisely the radiation pulse the next article constructs. So the present-centering is a testament to the machinery's consistency, and its breakdown under acceleration is not a bug but the chapter's climax: radiation is the correction notice for a failed extrapolation, delivered at the speed of light.

Why do we differentiate the LW potentials to get fields instead of using Jefimenko's equations directly?

Both routes are exact and they provably agree; the LW route wins on error control. Jefimenko's integrals applied to a point charge require collapsing deltas inside five separate terms — each with its own retarded-argument Jacobian bookkeeping — and then combining, a notorious factory for lost factors of $D$ . The LW route performs the treacherous collapse once, at the level of two scalar-valued potentials, and then differentiates — heavy but mechanical algebra, with the subtlety concentrated in known objects ( $\nabla t_r$ , $\partial t_r/\partial t$ , worked out next article). It is the standard division of labor met throughout this chapter: fields-first formulations for structure and doctrine, potential-first for computation. Historical footnote: Liénard and Wiechert both went through potentials; nobody has ever preferred the direct route for this problem, and the rare published attempts are cautionary tales.

Do the LW potentials satisfy the Lorenz condition, and does it matter?

Yes — they inherit it. The retarded potentials satisfy the Lorenz condition whenever the source conserves charge (the certificate derivation of the retarded-potentials article), and a point charge on a smooth trajectory conserves charge trivially (its $\rho$ and $\mathbf{J} = \rho\mathbf{v}$ obey continuity as distributions). One can also verify $\nabla\cdot\mathbf{A} + \mu_0\varepsilon_0\,\partial_t V = 0$ directly on (2) — a strenuous but instructive exercise in retarded differentiation, sometimes assigned as preparation for the next article's field computation. It matters for exactly one reason: the field formula $\mathbf{E} = -\nabla V - \partial_t\mathbf{A}$ presumes the pair lives in a single gauge, and the coming differentiation leans on that legality at every line. The chapter's interlock — continuity ⇒ gauge consistency ⇒ legal field assembly — runs through the point charge unbroken.

Is there an advanced Liénard-Wiechert solution too?

Certainly — run the identical collapse on the advanced Green function and everything goes through with the mirror bookkeeping: the advanced time $t_a$ (the unique future intersection of the forward light cone with the worldline), separation $\boldsymbol{\mathscr{r}}_a = \mathbf{r} - \mathbf{w}(t_a)$ , and Jacobian $(1 + \hat{\boldsymbol{\mathscr{r}}}_a\cdot\boldsymbol{\beta}_a)^{-1}$ — note the sign flip: an approaching charge is under-collected by the expanding sphere. The advanced LW potentials are exactly what Dirac's radiation-reaction construction (the advanced-potentials article) evaluates at the charge itself, and the half-difference of the two LW families, expanded about the present, is the classical route to the Abraham–Lorentz force at this chapter's end. So the advanced twin is not exotica here — it is scheduled infrastructure, and the sign-flipped Jacobian is worth deriving once yourself as a check that you own Derivation B rather than merely believing it.