Radiation Reaction Force

The chapter has been running an unbalanced ledger. The Larmor and Liénard formulas price the energy a charge broadcasts — but Newton's second law, as used so far, contains no term that collects that price from the particle. A charge shaken by an external force radiates real watts; if nothing debits its mechanical energy, we have built a perpetual-motion machine out of impeccable formulas. Something must push back on the charge, and the only candidate is the charge's own field. This article derives that self-force — the radiation reaction — and delivers the most notorious result in classical electrodynamics: $\mathbf{F}_{\text{rad}}=m\tau\dot{\mathbf{a}}$ , a force proportional not to position, not to velocity, not even to acceleration, but to the jerk, with the tiny built-in timescale $\tau=\mu_0q^2/6\pi mc\approx6.3\times10^{-24}$ s — two-thirds the light-crossing time of the classical electron radius. The derivation is a study in honest accounting. Energy balance fixes only the time-averaged work of the self-force, and converting the Larmor $a^2$ into a force costs an integration by parts whose boundary term — the Schott energy, a reservoir of energy parked in the near field — is the fine print that resolves every apparent paradox about "instantaneous" radiated power. A second route, Lorentz's model of a small charged shell pushing on itself with retarded fields, produces the same force plus a bonus that anticipates twentieth-century physics: a divergent self-energy term that must be absorbed into the observed mass — renormalization, discovered inside classical electrodynamics. What this article does not do is solve the resulting equation of motion. That is deliberate. The third-order equation harbors pathologies — runaway solutions, preacceleration — famous enough to earn their own article next, where they are prosecuted in full together with the modern (Landau–Lifshitz) resolution. Here we establish the force, its scales, and when it matters: almost never, and then overwhelmingly.

Key takeaways

Energy conservation demands a self-force: the Larmor power must be debited from the particle's mechanical energy by the charge's own field acting back on it.
Matching the time-averaged work to the radiated energy gives the Abraham–Lorentz force $\mathbf{F}_{\text{rad}}=\dfrac{\mu_0q^2}{6\pi c}\dot{\mathbf{a}}=m\tau\dot{\mathbf{a}}$ — proportional to the jerk.
The timescale $\tau=\mu_0q^2/6\pi mc=\tfrac23r_e/c\approx6.3\times10^{-24}$ s for the electron: radiation reaction matters only when the motion changes on that scale, i.e. essentially never below extreme-field frontiers.
The instantaneous identity $\mathbf{F}_{\text{rad}}\cdot\mathbf{v}=-P_{\text{Larmor}}+\tfrac{d}{dt}(m\tau\,\mathbf{v}\cdot\mathbf{a})$ shows the force pays the radiation bill only on average; the total-derivative piece is minus the rate of change of the Schott (acceleration) energy, lent and reclaimed by the near field.
Lorentz's charged-shell calculation confirms the force microscopically: a divergent term renormalizes the mass (with the notorious 4/3 bookkeeping defect); the finite, shell-independent remainder is exactly $m\tau\dot{\mathbf{a}}$ .
Applications: the classical linewidth of spectral lines ( $\gamma_{\text{cl}}=\omega_0^2\tau$ , the universal $1.2\times10^{-4}$ Å), radiation damping in storage rings, and the ultraintense-laser regime where the reaction force finally rivals the driving force.
Key idea: closing the chapter's energy books forces a jerk-proportional self-force with coefficient $m\tau$ — correct on average, third-order in time, and thereby the crack through which classical physics's pathologies enter (next article).

The core insight

Energy conservation forces the existence of a self-force; matching the cycle-averaged work to the Larmor power yields $\mathbf{F}_{\text{rad}}=m\tau\,\dot{\mathbf{a}}$ , a jerk-dependent force whose third-order character signals the limits of the classical point particle.

Where this lives today

For a force eight to fourteen orders below its rivals, radiation reaction runs a surprising amount of infrastructure. In every electron storage ring it is the design tool: synchrotron emission damps betatron and synchrotron oscillations (the covariant self-force, averaged over the lattice), shrinking injected beams into the micron-scale ribbons that light sources need — machines schedule "damping times" the way kitchens schedule ovens, and dedicated damping rings were core subsystems in linear-collider designs. The equilibrium between radiation damping and quantum emission noise sets the ultimate beam brightness — a working, daily-use balance of this article's force against its QED corrections. The extreme-field frontier has finally made the force a headline actor. Petawatt facilities colliding GeV electron beams with counter-propagating pulses reported, in 2018, the first laboratory signatures of radiation reaction beyond the perturbative regime — electron energy losses of tens of percent over the full pulse encounter, of order $40\ \mathrm{fs}$ , at a significance that does not yet separate classical (Landau–Lifshitz) from quantum-corrected models. Upcoming multi-10-PW machines (ELI, Apollon, and successors) are designed with radiation-reaction-dominated dynamics as a stated science goal: plasma physics where the self-force competes with the Lorentz force term by term. And conceptually, the shell calculation's legacy pervades physics: mass renormalization, effective-theory reasoning ("the point model is valid only above scale $\tau$ "), and reduction of order are all standard toolkit items whose first working appearance was a 1904 argument about a charged sphere disagreeing with its own past.

Origins · 1 minMax Abraham and how this was discovered

Max Abraham

1875–1922 · Göttingen — the theory's fiercest craftsman, wrong about relativity, right about the self-force

Abraham built the first systematic dynamics of the electron's own field. In 1902–03, extending Lorentz's work, he computed the electromagnetic momentum of a rigid charged sphere, derived the self-force including the $\dot{\mathbf{a}}$ term that now bears his and Lorentz's names, and championed the boldest idea of the pre-relativistic era: that all mass is electromagnetic — inertia as nothing but a charge's reluctance to disturb its own field. His "rigid electron" competed head-on with Lorentz's contractible one; the Kaufmann experiments of 1905–06 initially seemed to favor Abraham's velocity-dependence over the Lorentz–Einstein prediction, delaying relativity's acceptance by several years until better data reversed the verdict. Abraham never surrendered — he fought relativity in print with a technical brilliance even his opponents admired (his textbook on electromagnetic theory trained a generation, and Born called him a master of clarity), and later fought general relativity too, having constructed one of the first serious competitor gravity theories. He died at 47 of a brain tumor, a scientist on the losing side of the century's great turn whose self-force calculation, mass-renormalization logic, and 4/3 puzzle turned out to be permanent physics anyway.

History

1897

— Larmor's formula prices the radiation; the energy ledger opens

1902

— Abraham computes the rigid electron's self-force and electromagnetic momentum

1904

— Lorentz's shell expansion: divergent mass absorbed, finite jerk term isolated

1905

— Poincaré adds the binding stresses that later cure the 4/3 defect

1912

— Schott identifies the acceleration energy — the near-field ledger's lending account

1938

— Dirac's covariant point-charge analysis: the Abraham–Lorentz–Dirac equation

1951

— Landau and Lifshitz publish the reduction-of-order equation now standard in strong-field work

2018

— Gemini laser experiments report radiation-reaction signatures in single laser–electron collisions

The physical picture

Why must the force track the jerk? Start with what it cannot be. It cannot be a drag $-\lambda\mathbf{v}$ : uniform motion does not radiate, so a coasting charge must feel nothing. It cannot be $-\lambda\mathbf{a}$ either: that is indistinguishable from extra mass (it renormalizes $m$ and drains no energy on average over a cycle). The radiated power is $\propto a^2$ — quadratic, always negative in effect — but a force law must be linear in the motion if it is to come from the linear Maxwell equations. The resolution: linearity in $\dot{\mathbf{a}}$ , whose work $m\tau\dot{\mathbf{a}}\cdot\mathbf{v}$ integrates by parts to $-m\tau a^2$ plus boundary terms. The jerk is the lowest derivative that can pay a quadratic bill on average while remaining linear in the trajectory. The microscopic picture explains the same thing with retardation. Model the charge as a small ball of charge, radius $\varepsilon$ . Each element feels the retarded field of every other element — and when the ball accelerates, the geometry is lopsided: the field a trailing element receives from a leading one left when the ball was moving differently. Expand that lopsidedness in powers of $\varepsilon$ : the leading term scales as $1/\varepsilon$ and points along $-\mathbf{a}$ — an inertia-like drag that diverges for a point charge and gets absorbed into the measured mass. The next term is independent of $\varepsilon$ altogether, points along $+\dot{\mathbf{a}}$ , and survives the point limit with the universal coefficient $\mu_0q^2/6\pi c$ . The radiation reaction is what remains of a charge's argument with its own past after the infinite part has been hidden in the mass.

Figure

A charge arguing with its own retarded field — the shell model, and where each term of the self-force comes from

The scale $\tau$ deserves respect before any formula. It is the light-crossing time of (two-thirds of) the classical electron radius: the shortest time over which the notion "point charge with classical fields" can pretend to be consistent. Radiation reaction is a correction of relative size $\sim\omega\tau$ for motion at frequency $\omega$ — about $10^{-8}$ for optical light, $10^{-14}$ for gigahertz electronics. It is the smallest force in this book, until the moment (petawatt lasers, X-ray free-electron dynamics, ultrarelativistic storage rings) it is not.

The math: energy balance, and its fine print

The chapter's conservation laws demand that the work done by all forces on the charge account for both its kinetic-energy change and the energy shipped to infinity. Writing Newton's law as $m\mathbf{a}=\mathbf{F}_{\text{ext}}+\mathbf{F}_{\text{rad}}$ , the unknown $\mathbf{F}_{\text{rad}}$ must satisfy, over suitable intervals,

\int_{t_1}^{t_2}\mathbf{F}_{\text{rad}}\cdot\mathbf{v}\,dt=-\int_{t_1}^{t_2}\frac{\mu_0q^2a^2}{6\pi c}\,dt. \qquad (1)

where: the right side is minus the Larmor energy radiated (nonrelativistic regime; the covariant extension is deferred below). Note what (1) does and does not claim: it constrains an integral of the unknown force, not its instantaneous value — the entire logical character of what follows, including its limitations, descends from that distinction.

Derivation AThe Abraham-Lorentz force from energy balance

The goal is the classic result $\mathbf{F}_{\text{rad}}=\mu_0q^2\dot{\mathbf{a}}/6\pi c\equiv m\tau\dot{\mathbf{a}}$ , extracted from (1), with the boundary-term fine print kept in view rather than swept away — because that fine print (the Schott energy) is what later reconciles every instantaneous-bookkeeping puzzle. Strategy: integrate the $a^2$ by parts to shift a derivative off one factor of acceleration and onto the velocity, identify conditions killing the boundary term, then read off the integrand. The move is one line; the physics is in the honesty about when the line is valid.

Step 1. Integrate the radiated energy by parts. With $\mathbf{a}=\dot{\mathbf{v}}$ :

\int_{t_1}^{t_2}\mathbf{a}\cdot\mathbf{a}\,dt=\Big[\mathbf{a}\cdot\mathbf{v}\Big]_{t_1}^{t_2}-\int_{t_1}^{t_2}\dot{\mathbf{a}}\cdot\mathbf{v}\,dt.

Step 2. Kill the boundary term — and record the price. For periodic motion (endpoints one full cycle apart), or whenever the state returns $\mathbf{a}\cdot\mathbf{v}$ to its initial value, the bracket vanishes. Then (1) becomes

\int_{t_1}^{t_2}\mathbf{F}_{\text{rad}}\cdot\mathbf{v}\,dt=\frac{\mu_0q^2}{6\pi c}\int_{t_1}^{t_2}\dot{\mathbf{a}}\cdot\mathbf{v}\,dt.

Why the price must be recorded: for non-periodic stretches the bracket does NOT vanish, and the self-force's work differs from the radiated energy by exactly that boundary quantity. The derivation is therefore silent about any force component whose work integrates to zero on the chosen intervals — the identification below is the simplest candidate consistent with (1), spectacularly confirmed by the independent shell calculation, not a unique logical consequence.

Step 3. Identify integrands:

\mathbf{F}_{\text{rad}}=\frac{\mu_0q^2}{6\pi c}\,\dot{\mathbf{a}}=m\tau\,\dot{\mathbf{a}},\qquad \tau\equiv\frac{\mu_0q^2}{6\pi mc}=\frac{2}{3}\frac{r_e}{c}\approx6.3\times10^{-24}\ \text{s (electron)}.

Step 4. The instantaneous ledger — the Schott identity. Undo the averaging to see what the force does moment by moment:

\mathbf{F}_{\text{rad}}\cdot\mathbf{v}=m\tau\,\dot{\mathbf{a}}\cdot\mathbf{v}=-\underbrace{\frac{\mu_0q^2a^2}{6\pi c}}_{P_{\text{Larmor}}}+\frac{d}{dt}\Big(m\tau\,\mathbf{v}\cdot\mathbf{a}\Big).

Why this identity is the article's quiet centerpiece: instantaneously, the self-force's work does NOT equal minus the radiated power — the difference is MINUS the rate of change of the Schott energy $E_S=-m\tau\,\mathbf{v}\cdot\mathbf{a}$ , a reservoir carried in the charge's near field that lends energy during parts of a cycle and reclaims it in others. Averaged over a period the loan nets to zero and the Larmor bill stands; on shorter stretches the near field is an active party to the books. Every classic puzzle — "uniform acceleration radiates but feels no jerk force; who pays?" — is settled by this term: the Schott reservoir pays, continuously, for as long as the uniform acceleration lasts. *Note: the Gaussian form of the same coefficient is $2q^2\dot a/3c^3$ ; dimensionally $[m\tau][\dot a]=$ kg·s·m/s³ = N, confirming the force units; and the harmonic-oscillator average of the Schott term vanishes exactly, reconciling the instantaneous identity with the cycle-averaged Larmor balance.*

Go deeper · Derivation B · 2 minThe Lorentz shell — mass renormalization and the universal finite term

Derivation BThe Lorentz shell — mass renormalization and the universal finite term

The goal is the microscopic confirmation: computing the net force a small, rigid, spherical shell of charge (radius $\varepsilon$ , total charge $q$ ) exerts on itself through retarded fields, and extracting the two leading terms — the divergent electromagnetic mass and the finite Abraham–Lorentz force. The strategy: pair up charge elements, expand each retarded interaction in the small delay $d/c$ (element separation $d\lesssim2\varepsilon$ ), and sum. We run the expansion at structure level — the tensor bookkeeping of the exact sphere average is Jackson §16.3's page of algebra — because the two physics conclusions (divergence renormalized; finite term universal) are what the rest of the story stands on.

Step 1. Set up the pairwise retarded force. The net self-force is $\mathbf{F}_{\text{self}}=\int\!\!\int\rho(\mathbf{r})\,\mathbf{E}_{\text{ret}}[\rho(\mathbf{r}')]\,d^3r\,d^3r'$ — each element in the field of every other. For a rigid nonrelativistic shell, the only asymmetry between a pair's mutual pushes comes from retardation: the news each receives left when the shell moved differently.

Step 2. Expand in the retardation. Taylor-expanding the source's position and the field in the delay $d/c$ generates a series in powers of $d$ (hence of $\varepsilon$ ) with alternating structure:

\mathbf{F}_{\text{self}}=-\,\delta m\,\mathbf{a}\;+\;\frac{\mu_0q^2}{6\pi c}\,\dot{\mathbf{a}}\;+\;O(\varepsilon),\qquad \delta m=\frac{4}{3}\,\frac{U_{\text{es}}}{c^2},\quad U_{\text{es}}=\frac{q^2}{8\pi\varepsilon_0\varepsilon}.

Why the first term diverges and what to do with it: $U_{\text{es}}$ is the shell's electrostatic self-energy, blowing up as $\varepsilon\to0$ . But a force $\propto-\mathbf{a}$ is operationally indistinguishable from inertia: move it to the left of Newton's law and it adds to the bare mass. The observed mass is $m=m_{\text{bare}}+\delta m$ , and only $m$ is measurable — mass renormalization, executed in 1904 with the same logic (subtract a divergence into an unobservable bare parameter) that quantum field theory would industrialize forty years later.

Step 3. Note the scandal in the coefficient. The $4/3$ is wrong: relativity demands a field energy $U_{\text{es}}$ contribute mass $U_{\text{es}}/c^2$ , not $\tfrac43U_{\text{es}}/c^2$ . The defect is real but diagnosable — a rigid charged shell is not a closed system (something non-electromagnetic must hold it against its own repulsion: the Poincaré stresses), and a covariant accounting that includes the binding agent restores the factor to $1$ . The lesson is honest bookkeeping, not broken electrodynamics — and the whole complex of issues is the advanced self-energy article's subject.

Step 4. Read the finite term. The $O(\varepsilon^0)$ term carries no memory of $\varepsilon$ , none of the shell's shape (any charge distribution gives the same coefficient), and survives the point limit intact: it is a universal property of a charge interacting with its own retardation, and it is exactly the force Derivation A extracted from energy balance. Two derivations, opposite starting points — global bookkeeping versus microscopic mechanics — one force.

Assembling Newton's law with the new force gives the equation this pair of articles pivots on:

m\,\mathbf{a}=\mathbf{F}_{\text{ext}}+m\tau\,\dot{\mathbf{a}}. \qquad (2)

where: the innocuous-looking $m\tau\dot{\mathbf{a}}$ raises the order: (2) is third-order in position, so position and velocity no longer determine the future — a third datum is needed, and the extra solution family it unlocks contains exponential runaways and acausal preacceleration. Stating, solving, and defusing (2) — Griffiths's "ointment in the fly" — is the next article's entire business; this one closes with where the force legitimately rules.

Figure

The Schott ledger — instantaneous books need three parties: particle, near field, and radiation

Limiting cases

case	result	meaning
$\omega\tau\ll1$ (all ordinary physics)	$F_{\text{rad}}/F_{\text{ext}}\sim\omega\tau$	a perturbation: $10^{-8}$ at optical frequencies, $10^{-14}$ at GHz
periodic motion	$\langle\mathbf{F}_{\text{rad}}\cdot\mathbf{v}\rangle=-\langle P_{\text{Larmor}}\rangle$	the averaged books balance exactly; Schott term nets to zero
uniform acceleration $\dot{\mathbf{a}}=0$	$\mathbf{F}_{\text{rad}}=0$ yet $P>0$	the Schott reservoir pays; no paradox, just near-field lending
bound electron, frequency $\omega_0$	damping rate $\gamma_{\text{cl}}=\omega_0^2\tau$	classical linewidth; the universal $\Delta\lambda=2\pi c\tau=1.2\times10^{-4}$ Å
free particle ( $\mathbf{F}_{\text{ext}}=0$ ) in (2)	$\mathbf{a}=\mathbf{a}_0e^{t/\tau}$ exists	the runaway mode — prosecuted next article
relativistic completion	Abraham–Lorentz–Dirac: $F^\mu=\frac{\mu_0q^2}{6\pi c}\big(\ddot{u}^\mu+\frac{u^\mu}{c^2}\dot{u}^\nu\dot{u}_\nu\big)$	the covariant self-force; the extra term keeps $u^\mu u_\mu$ fixed — advanced article
$\hbar$ turned on	QED: photon recoil, vertex corrections	the classical crisis dissolves into quantum bookkeeping below $\sim\tau$

Common trapThe force does not equal minus the radiated power divided by velocity — instantaneous energy balance is the wrong demand

The single most common misreading of radiation reaction: treating $\mathbf{F}_{\text{rad}}\cdot\mathbf{v}=-P_{\text{Larmor}}$ as an instantaneous identity and "deriving" nonsense from it (a charge momentarily at rest with $a\ne0$ radiates but no candidate force can do work — contradiction?). The energy-balance derivation never claimed instantaneous matching: it fixed the force only through interval-averaged work, and the honest instantaneous statement is the Schott identity, $\mathbf{F}_{\text{rad}}\cdot\mathbf{v}=-P+\frac{d}{dt}(m\tau\mathbf{v}\cdot\mathbf{a})$ . The near field is a co-signatory on the account — it lends during one phase of the motion and collects during another. Corollary of the same fine print: since energy balance constrains only integrals, the force law is fixed only up to terms whose work vanishes on the qualifying intervals; its ultimate justification is the shell calculation and, at the relativistic level, the Dirac analysis. Quote the pair — averaged Larmor debit plus Schott ledger — or quote wrong physics.

Worked examples

1 · How small is small — the reaction force across the applications (confidence builder)

Setup. Compare $F_{\text{rad}}$ to the driving force for an electron oscillating (i) in a 500 nm optical field and (ii) in a 1 GHz antenna. Then find the regime where the ratio reaches unity.

Step 1. The ratio for harmonic motion. With $\mathbf{v}=v_0\cos\omega t$ : $|\mathbf{F}_{\text{rad}}|=m\tau|\dot{\mathbf{a}}|=m\tau\omega^2|v|$ ... against the inertial scale $m|\mathbf{a}|=m\omega|v|$ :

\frac{F_{\text{rad}}}{ma}=\omega\tau.

One dimensionless number decides everything.

Step 2. Optical: $\omega=3.8\times10^{15}$ s⁻¹, so $\omega\tau=(3.8\times10^{15})(6.3\times10^{-24})=\mathbf{2.4\times10^{-8}}$ . Even for light, radiation reaction is a twenty-parts-per-billion correction per cycle.

Step 3. Radio: $\omega=6.3\times10^9$ s⁻¹ gives $\omega\tau=\mathbf{3.9\times10^{-14}}$ . Antenna engineers never meet this force directly — they meet its integrated alias, radiation resistance, which is the same physics booked at the circuit level.

Step 4. Where does $\omega\tau\to1$ live? $\omega\sim1/\tau=1.6\times10^{23}$ s⁻¹ — photon energies $\hbar\omega\sim100$ MeV, far beyond electron rest mass: the classical crossover point lies deep inside QED's jurisdiction, which is the cleanest statement that classical radiation reaction is always either negligible or already quantum. The loophole: relativistic boosts. In the electron's rest frame an oncoming petawatt laser is blueshifted and intensified by $\sim\gamma$ -dependent factors, and modern experiments (GeV electrons colliding with ultraintense pulses) push the effective ratio to order unity — the one place the classical force becomes a leading actor just as its quantum replacement auditions.

The lesson: compute $\omega\tau$ first; it tells you whether radiation reaction is a footnote ( $<10^{-6}$ , almost always) or the story.

2 · The classical linewidth — radiation damping as spectroscopy (exam level)

Setup. Model an atomic electron as a classical oscillator of natural frequency $\omega_0$ ( $\lambda=500$ nm). Include radiation reaction, find the damping rate, the oscillator's $Q$ , and the resulting spectral linewidth.

Step 1. The damped equation. For bound motion at frequency $\omega_0$ , the reaction force $m\tau\dot{\mathbf{a}}$ evaluates on-shell as $m\tau\dot{\mathbf{a}}\approx-m\tau\omega_0^2\mathbf{v}$ (each derivative of harmonic motion brings $\omega_0$ , with $\dot{\mathbf{a}}=-\omega_0^2\mathbf{v}$ exactly for pure oscillation): a genuine velocity drag,

\ddot{x}+\gamma_{\text{cl}}\,\dot{x}+\omega_0^2x=0,\qquad \gamma_{\text{cl}}=\omega_0^2\tau.

(This substitution — using the undamped motion inside the tiny reaction term — is the reduction-of-order move that the next article elevates into the Landau–Lifshitz equation.)

Step 2. Numbers. $\omega_0=3.8\times10^{15}$ s⁻¹: $\gamma_{\text{cl}}=(3.8\times10^{15})^2(6.3\times10^{-24})=8.9\times10^{7}$ s⁻¹ — an energy decay time $1/\gamma_{\text{cl}}=\mathbf{11\ ns}$ . The oscillator's quality factor: $Q=\omega_0/\gamma_{\text{cl}}=1/\omega_0\tau=\mathbf{4.2\times10^7}$ .

Step 3. The linewidth. A decaying oscillation radiates a Lorentzian line of FWHM $\Delta\omega=\gamma_{\text{cl}}$ ; in wavelength,

\Delta\lambda=\frac{\lambda^2}{2\pi c}\,\Delta\omega=\frac{\lambda^2}{2\pi c}\,\omega_0^2\tau=2\pi c\,\tau=\mathbf{1.2\times10^{-4}\ Å},

independent of $\lambda$ : every classical line, whatever its color, has the same natural width — a pure constant of the electron, $2\pi c\tau=4\pi r_e/3$ .

Step 4. Confront reality. Measured allowed transitions: lifetimes of nanoseconds (hydrogen $2p$ : 1.6 ns), Lorentzian shapes, natural widths within factors of a few of $1.2\times10^{-4}$ Å. The classical oscillator gets the lineshape exactly and the scale remarkably — because the quantum result for a strong transition is the classical rate times an oscillator strength of order unity. This was the evidence that radiation damping is real physics, decades before anyone could see $\tau$ directly.

The lesson: radiation reaction's most successful prediction is not a force measurement but a spectral shape: the Lorentzian natural linewidth, with its universal classical width — the fingerprint of a charge taxed by its own field.

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., §11.2.2–11.2.3 · Jackson, Classical Electrodynamics, 3rd ed., §16.1–16.3 · Landau & Lifshitz, The Classical Theory of Fields, §75–76 · Rohrlich, Classical Charged Particles, 3rd ed. · Zangwill, Modern Electrodynamics, §23.7 Key numerical anchors: $\tau=6.266\times10^{-24}$ s (consistent from both $\tau=\mu_0q^2/6\pi mc$ and $\tau=\tfrac23r_e/c$ with $r_e=2.818\times10^{-15}$ m); $\omega\tau=2.4\times10^{-8}$ at optical frequencies, $3.9\times10^{-14}$ at 1 GHz; $\gamma_{\text{cl}}=8.9\times10^7$ s⁻¹ (decay time 11 ns, $Q=4.2\times10^7$ ) for the 500 nm oscillator; $\Delta\lambda=2\pi c\tau=1.18\times10^{-4}$ Å for the classical natural linewidth; shell electromagnetic mass $\delta m/m=\tfrac23$ at $\varepsilon=r_e$ .

Builds on: Larmor Formula · Relativistic Larmor Formula (Liénard's Result) · Fields of a Moving Point Charge · Liénard–Wiechert Potentials Leads to: Abraham–Lorentz Equation and Its Pathologies · Advanced Potentials and Causality

If the self-force comes from the charge's own field, why is not the whole Coulomb field already included in Newton's law?

Because the symmetric part of the self-interaction cancels or renames itself. A static charge's field pulls on the charge equally in all directions: zero net force. When the charge accelerates, the retarded asymmetry produces the two survivors of Derivation B: a piece proportional to $-\mathbf{a}$ , which is physically inertia (you cannot distinguish "hard to accelerate because massive" from "hard to accelerate because the field resists rearrangement") and is absorbed into the one mass you actually measure; and the finite jerk piece, which alone does averaged work and alone is called the radiation reaction. So the Coulomb self-field is included — partly as your mass, partly as this force. What was never included, and never will be classically, is a finite prescription for the point limit's divergent self-energy: that bill goes to quantum field theory.

Is the Schott energy real energy, or an accounting fiction?

Real, locatable, and occasionally spendable. The total field energy of an accelerating charge splits into the far-zone radiation (irreversibly gone) and near-zone contributions bound to the charge; the Schott term $E_S=-m\tau\,\mathbf{v}\cdot\mathbf{a}$ is the velocity-acceleration cross term of that bound energy — it lives in the overlap between the Coulomb-like field and the acceleration field, within roughly a wavelength of the particle. It changes sign with the motion's phase (negative when speeding up along $\mathbf{v}$ , positive when braking), which is exactly the lending behavior: during uniform acceleration it drains at precisely the Larmor rate, paying for radiation the jerk force cannot. It is "fictional" only in the sense that any energy partition of a single total field is conventional; its rate of change enters the observable bookkeeping unambiguously. Schott identified it in 1912, and it remains the standard resolution of the uniform-acceleration and hyperbolic-motion puzzles.

Does radiation reaction violate Newton's third law — what does the charge push back on?

The field — and the field is a full mechanical citizen, so no violation. The conservation chapter established that the electromagnetic field carries momentum and energy of its own; Newton's third law in its naive matter-on-matter form fails in electrodynamics generally, replaced by total (matter + field) momentum conservation. Radiation reaction is that theorem applied to a charge and its own emissions: the outgoing wave carries momentum (for relativistic motion, strongly beamed forward), and the particle recoils. The nonrelativistic jerk force is the low-velocity residue of this recoil bookkeeping; covariantly, the emitted four-momentum flows at rate $P u^\mu/c^2$ and the ALD self-force contains exactly the compensating term. The "reaction partner" is not a distant absorber (though Wheeler–Feynman theory explored crediting it there — see the advanced-potentials article) but the field the charge just created.

Why do protons and heavy ions effectively never feel radiation reaction?

Two mass penalties compound. First, at equal force a heavy particle accelerates less: Larmor power $\propto a^2\propto1/m^2$ . Second, the reaction timescale itself shrinks: $\tau\propto1/m$ , so $\tau_p=\tau_e(m_e/m_p)=3.4\times10^{-27}$ s and the dimensionless dial $\omega\tau$ is a further $1836$ times smaller for the same motion. Combined with the $\gamma^4/m^4$ scaling of synchrotron loss at fixed energy, the practical statement is stark: LHC protons at 7 TeV lose keV per turn (radiation damping time: about a day, actually used for gentle beam cooling), while equivalent-energy electrons would be unstorable. Radiation reaction is, to an excellent approximation, an electron-only phenomenon — and via the $q^4$ scaling of the damping, a strong argument for why heavy-ion machines ignore it entirely.

How does the quantum theory replace this force — is there a jerk term in QED?

No jerk term survives; the mechanism is reassembled from discrete recoils. In QED a radiating electron emits photons probabilistically; each emission kicks the electron's momentum, and the average momentum flow reproduces the classical (Landau–Lifshitz form of the) radiation reaction in the appropriate limit — with quantum corrections controlled by the parameter $\chi$ (the field strength in the electron's rest frame measured against the Schwinger critical field). Classical radiation reaction is thus the coherent, many-photon limit of recoil; the pathologies of the point-charge equation never arise because QED regularizes exactly the short-distance physics the classical shell model parameterized by $\varepsilon$ — with renormalization, the shell's own trick, now systematic. The modern strong-field experiments are precisely mapping the handover: classical Landau–Lifshitz at $\chi\ll1$ , stochastic photon-recoil corrections as $\chi\to1$ . The classical force is not wrong; it is the smooth shadow of a shot-noise process.

Given all the fine print, what exactly is the domain where the Abraham-Lorentz force is straightforwardly correct?

A comfortable and well-marked one: nonrelativistic motion whose timescales are all long compared to $\tau$ , treated to first order in $\tau$ , with averaged energy bookkeeping. There the force is unambiguous, its predictions (linewidths, damping rates, antenna radiation resistance, storage-ring damping in covariant extension) are quantitatively confirmed, and none of the pathologies activate — they live entirely in taking the third-order equation literally on timescales $\sim\tau$ , outside the theory's warranty. The honest hierarchy: (i) $\omega\tau\ll1$ , perturbative — use it freely, preferably in reduced (Landau–Lifshitz) form; (ii) $\omega\tau\sim1$ classically — the model is announcing its own breakdown; consult QED; (iii) the literal third-order dynamics — a fascinating pathology lab, the next article, valuable precisely as a case study in how a good theory fails at its edge.