Synchrotron and Inverse-Compton Radiation

The most luminous objects in the universe do not shine as stars do. Quasars, blazars, radio galaxies, and supernova-remnant shells pour out a nonthermal continuum that runs unbroken from metre-wavelength radio to teraelectronvolt gamma rays — a spectrum no blackbody can make. Two processes, acting on the same population of relativistic electrons, make nearly all of it. When such an electron spirals in a magnetic field it is accelerated centripetally and radiates: synchrotron emission, the ultrarelativistic descendant of cyclotron radiation. When the same electron strikes an ambient photon and hands it energy, the photon is boosted to far higher frequency: inverse-Compton scattering, ordinary Compton scattering run in reverse, with the electron rather than the photon as the energetic partner. The reason to treat them together is that they are, at bottom, the same physics. Both are Thomson scattering of a field off a fast electron — the virtual photons of a static magnetic field in one case, real ambient photons in the other — so both radiate the identical power $P=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U$ , differing only in whether $U$ is the magnetic energy density $U_B$ or the radiation energy density $U_{\rm rad}$ . That single fact is a gift to observers: the ratio of the two luminosities is simply $U_{\rm rad}/U_B$ , turning a pair of measured spectral humps into a direct reading of a cosmic magnetic field. Feed either process a power-law distribution of electron energies — the generic product of shock acceleration — and out comes a power-law photon spectrum whose slope encodes the acceleration microphysics. This article builds that machinery from the single electron up to the double-humped blazar.

Key takeaways

Synchrotron radiation is emitted by relativistic electrons spiralling in magnetic fields; the single-electron spectrum is broadband and peaks near a critical frequency $\nu_c\propto\gamma^2 B$ .
Inverse-Compton scattering lets the same electrons upscatter soft seed photons, boosting each photon's energy by a factor of order $\gamma^2$ — a microwave photon becomes an X-ray.
Both processes radiate the identical single-electron power $P=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U$ , so their luminosity ratio is $L_{\rm IC}/L_{\rm sync}=U_{\rm rad}/U_B$ — a clean magnetometer for distant sources.
A power-law electron spectrum $N(\gamma)\propto\gamma^{-p}$ produces a power-law photon spectrum $S_\nu\propto\nu^{-\alpha}$ with $\alpha=(p-1)/2$ , the defining signature of nonthermal emission.
Because power grows as $\gamma^2$ but energy only as $\gamma$ , high-energy electrons cool first ( $t_{\rm cool}\propto1/\gamma$ ), imprinting a spectral break of $\Delta\alpha=0.5$ ; runaway internal-Compton losses cap the brightness temperature near $10^{12}$ K, the Compton catastrophe.
Key idea: synchrotron and inverse-Compton are one process — Thomson scattering of a field off a relativistic electron — each radiating $P=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U$ with $U=U_B$ or $U_{\rm rad}$ , so their luminosity ratio is $U_{\rm rad}/U_B$ and both turn an $N(\gamma)\propto\gamma^{-p}$ electron spectrum into an $S_\nu\propto\nu^{-(p-1)/2}$ photon spectrum.

The core insight

Synchrotron and inverse-Compton are the same scattering of a field off a relativistic electron; each radiates $P=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U$ with $U=U_B$ for the magnetic field or $U=U_{\rm rad}$ for a photon bath, so their luminosity ratio equals $U_{\rm rad}/U_B$ and both convert an $N(\gamma)\propto\gamma^{-p}$ electron spectrum into an $S_\nu\propto\nu^{-(p-1)/2}$ photon spectrum.

Where this lives today

The two processes are the daily working language of high-energy astrophysics. The Fermi Large Area Telescope has catalogued thousands of blazars whose gamma-ray hump is inverse-Compton emission, and the Cherenkov arrays — H.E.S.S., MAGIC, VERITAS, and now CTA — map that hump up to tens of teraelectronvolts, sharpening the leptonic-versus-hadronic debate over whether the highest-energy photons are upscattered by electrons or made by protons. Synchrotron-self-Compton and external-Compton models are fit routinely to blazar spectral energy distributions to extract the jet's field, Doppler factor, and electron spectrum; the 2017 IceCube neutrino from the flaring blazar TXS 0506+056 hinted that some carry a hadronic component too. In the radio lobes, Chandra detects the inverse-Compton X-ray glow of electrons scattering the CMB, delivering the field measurements of Worked Example 2, while the Event Horizon Telescope's polarized image of the M87 jet base traces the synchrotron plasma at the launch point. The Compton catastrophe — the $\sim10^{12}$ K ceiling above which internal inverse-Compton losses run away — is why blazar cores that appear to exceed it are read as Doppler-beamed toward us.

The physical picture

Picture a relativistic electron threading a helix around a magnetic field line. At every instant it is accelerated toward the axis of its spiral, and — this is the whole story of synchrotron — relativistic beaming crushes all of its radiation into a narrow cone of half-angle about $1/\gamma$ along its velocity. As the electron gyrates, that cone sweeps around like a lighthouse beam, and an observer catches it for only a sliver of each orbit, so what arrives is a train of sharp pulses. A short pulse is rich in high harmonics, and two factors of $\gamma$ shorten it: one from beaming the emission into $1/\gamma$ of the orbit, one from the emitting point racing toward the observer at nearly $c$ , telescoping the pulse in arrival time. The two multiply, which is why the characteristic frequency scales as $\gamma^2 B$ .

Figure

Synchrotron beaming: an electron gyrating in a magnetic field B radiates into a forward cone of half-angle about 1 over gamma, so an observer sees a train of brief flashes whose shortness sets the critical frequency ν_c = 1.5 γ² ν_g sinα — for γ = 10⁴ in B = 10 μG this lands at ν_c ≈ 4 GHz, in the radio

Inverse-Compton is the same electron viewed against a sea of soft photons instead of a magnetic field. Sit in the electron's rest frame: incoming photons are blueshifted by a factor of order $\gamma$ , scatter in the ordinary Thomson way with almost no change in their rest-frame energy, and are blueshifted again by another factor of $\gamma$ on the way back to the laboratory. Two boosts, two factors of $\gamma$ , a net gain of order $\gamma^2$ — a 2.7-kelvin microwave photon struck by a $\gamma=10^4$ electron leaves as a hard X-ray. The deep point is that a static magnetic field, through its Lorentz force, acts on the electron exactly as a photon field of energy density $U_B=B^2/2\mu_0$ would, so the two are faces of one process, and dividing their powers hands the observer $U_{\rm rad}/U_B$ directly. And because shock acceleration naturally makes a power law of electron energies, both mechanisms print a power-law photon spectrum whose slope carries the acceleration physics across the cosmos.

Figure

The blazar double hump: the same electrons make a low-energy synchrotron bump and a high-energy inverse-Compton bump, and the ratio of the areas under them equals U_rad over U_B, turning two measured peaks into a magnetic-field measurement; the optically-thin falloff has Fν ∝ ν^−α with α = (p−1)/2 = 0.75 for p = 2.5

The core equations

The whole subject rests on one power law for a single electron. Because power is a Lorentz invariant, we compute it in the electron's rest frame — just Larmor's formula — and the result, expressed with the Thomson cross section, is identical for a magnetic field and a photon bath once we identify the right energy density.

P=\frac{4}{3}\,\sigma_T\,c\,\gamma^2\beta^2\,U \qquad (1)

where: $P$ — total radiated power of one electron; $\sigma_T=6.652\times10^{-29}\ \mathrm{m^2}$ — Thomson cross section; $\gamma$ — Lorentz factor; $\beta=v/c$ ; $U$ — field energy density, $U_B=B^2/2\mu_0$ for synchrotron or $U_{\rm rad}$ for inverse-Compton. Reading it: power climbs as $\gamma^2$ and linearly in the field energy density. A single electron does not radiate at one frequency: Fourier-analysing its pulse train spreads the power over a broad continuum peaking near a critical frequency — the gyrofrequency lifted by the two factors of $\gamma$ above and modulated by the pitch angle.

\nu_c=\frac{3}{4\pi}\,\gamma^2\,\frac{eB\sin\alpha}{m_e c}=\frac{3}{2}\,\gamma^2\,\nu_g\sin\alpha \qquad (2)

where: $\nu_c$ — critical frequency where the single-electron spectrum peaks; $\nu_g=eB/(2\pi m_e c)$ — nonrelativistic gyrofrequency; $\alpha$ — pitch angle; $e,m_e$ — electron charge and mass. Reading it: the $\gamma^2$ scaling lets gigahertz radio report electrons with $\gamma\sim10^4$ in microgauss fields. Dividing equation (1) for the two field cases makes the central diagnostic fall out with no extra physics: the same electrons deliver synchrotron and inverse-Compton luminosities in the ratio of the two energy densities.

\frac{L_{\rm IC}}{L_{\rm sync}}=\frac{U_{\rm rad}}{U_B} \qquad (3)

where: $U_{\rm rad}$ — energy density of the ambient (seed) photon field; $U_B=B^2/2\mu_0$ — magnetic energy density. Reading it: measure both humps and you have measured the field, given the seed field — which for extended lobes is the precisely known cosmic microwave background. Real sources hold a distribution of electrons, and shock acceleration makes it a power law in energy. Convolving that with the single-electron emission — each electron radiating near $\nu\propto\gamma^2$ — turns a power law in $\gamma$ into a power law in $\nu$ , the fingerprint of nonthermal emission.

N(\gamma)\,\mathrm d\gamma\propto\gamma^{-p}\,\mathrm d\gamma \;\Longrightarrow\; S_\nu\propto\nu^{-\alpha},\qquad \alpha=\frac{p-1}{2} \qquad (4)

where: $N(\gamma)$ — electrons per unit Lorentz factor; $p$ — electron energy index; $S_\nu$ — flux density; $\alpha$ — spectral index. Reading it: a measured $\alpha$ gives $p=2\alpha+1$ ; the shock value $p\approx2$ gives $\alpha\approx0.5$ , and cooling steepens it toward $\alpha\approx1$ . For inverse-Compton in the Thomson regime the mean scattered-photon energy is boosted by the same $\gamma^2$ that shortened the synchrotron pulse, carrying the identical coefficient $4/3$ as the power — a coincidence that is really the shared origin speaking.

\langle E_\gamma^{\rm out}\rangle=\frac{4}{3}\,\gamma^2\,E_\gamma^{\rm in}\qquad(\gamma E_\gamma^{\rm in}\ll m_e c^2) \qquad (5)

where: $E_\gamma^{\rm in}$ — seed-photon energy; $E_\gamma^{\rm out}$ — scattered-photon energy; the bracket is an average over isotropic directions. Reading it: for CMB seeds ( $E_\gamma^{\rm in}\sim6\times10^{-4}$ eV) and $\gamma\sim10^4$ the output is an X-ray — why radio-lobe electrons reappear in Chandra images. Finally, because power scales as $\gamma^2$ while energy scales only as $\gamma$ , the radiative lifetime shrinks with energy — the fastest electrons die first, the clock behind every cooling break.

t_{\rm cool}=\frac{E}{P}=\frac{\gamma m_e c^2}{\tfrac{4}{3}\sigma_T c\,\gamma^2 U}=\frac{3\,m_e c}{4\,\sigma_T\,\gamma\,U} \qquad (6)

where: $t_{\rm cool}$ — radiative lifetime; $E=\gamma m_e c^2$ — electron energy; $U=U_B+U_{\rm rad}$ , since both channels cool the electron. Reading it: $t_{\rm cool}\propto1/(\gamma U)$ , so the highest- $\gamma$ electrons in the strongest fields vanish soonest, carving a break where cooling time equals source age.

Derivation AWhy synchrotron and inverse-Compton radiate the same power

The goal is to show, from the relativistic Larmor formula, that an electron accelerated by a magnetic field and one scattering an isotropic photon bath radiate the identical power $P=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U$ , so the distinction between the two mechanisms collapses to whether the driving energy density is magnetic or radiative — and dividing the two cases yields the luminosity ratio $U_{\rm rad}/U_B$ with no further assumptions.

Step 1. Start in the electron's instantaneous rest frame, where nonrelativistic Larmor gives $P'=e^2 a'^2/(6\pi\varepsilon_0 c^3)$ . Total radiated power is a Lorentz invariant, $P=P'$ , so we need only the rest-frame acceleration. Why: invariance lets us do the hard relativistic kinematics once, on the acceleration, not on the radiation field.

Step 2. In the magnetic case the transverse Lorentz force in the rest frame is $F'_\perp=e\gamma\beta cB\sin\alpha$ , so $a'_\perp=e\gamma\beta cB\sin\alpha/m_e$ ; the factor of $\gamma$ from boosting the field, squared inside $a'^2$ , produces the $\gamma^2$ in the power. Why: the entire energy dependence of synchrotron traces to one transformation of the magnetic force.

Step 3. Substitute $a'$ into Larmor and average over isotropic pitch angles with $\langle\sin^2\alpha\rangle=2/3$ . Regrouping the constants with $\sigma_T=e^4/(6\pi\varepsilon_0^2 m_e^2 c^4)$ gives $P_{\rm sync}=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2\,(B^2/2\mu_0)=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U_B$ . Why: recognizing $B^2/2\mu_0$ as the magnetic energy density reveals the field, not the field strength, as the relevant quantity, and the pitch-angle average is what sets the coefficient to $4/3$ .

Step 4. Repeat with an isotropic bath of real photons of energy density $U_{\rm rad}$ : averaging the Thomson-scattered power over incoming directions reproduces $P_{\rm IC}=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U_{\rm rad}$ , the same expression with $U_B\to U_{\rm rad}$ . Dividing gives $L_{\rm IC}/L_{\rm sync}=U_{\rm rad}/U_B$ ; a check confirms $[\sigma_T c\,U]=\mathrm{W}$ . Why: this is the unification — a magnetic field and a photon field enter the radiation identically, so the luminosity ratio is a pure field-energy ratio.

Go deeper · Derivation B · 1 minThe two-boost origin of the gamma-squared energy gain

Derivation BThe two-boost origin of the gamma-squared energy gain

The aim is to trace the $\gamma^2$ amplification of inverse-Compton scattering to its source: two successive Lorentz boosts, into the electron rest frame and back out, each contributing a Doppler factor of order $\gamma$ , with near-elastic Thomson scattering between — and to fix the coefficient $4/3$ by averaging over isotropic directions while flagging where the clean scaling breaks down.

Step 1. Boost a seed photon of lab energy $E_\gamma$ into the electron rest frame: $E'=\gamma(1+\beta\cos\theta)E_\gamma$ , about $2\gamma E_\gamma$ for a head-on encounter — one factor of $\gamma$ up. Why: in the rest frame the electron is at rest, so the collision is just Thomson scattering — but only after the photon has been blueshifted.

Step 2. Scatter in the rest frame. Provided $E'\ll m_e c^2$ , the scattering is Thomson: the photon energy is essentially unchanged, $E''\approx E'$ , and only its direction is randomized. Why: this elastic-in-the-rest-frame condition is the entire content of the Thomson limit — it keeps the process clean and linear in $E_\gamma$ .

Step 3. Boost back to the laboratory: $E_\gamma^{\rm out}=\gamma(1+\beta\cos\theta')E''$ , supplying the second factor of $\gamma$ , so a typical outgoing photon carries $\sim\gamma^2$ times the seed energy. Why: the symmetry between the two boosts is why the gain is $\gamma^2$ and not $\gamma$ — blueshifted on the way in and again on the way out.

Step 4. Average over isotropic incoming and outgoing directions: the angular integrals replace the head-on factors of 2 by their mean and deliver $\langle E_\gamma^{\rm out}\rangle=\tfrac{4}{3}\gamma^2 E_\gamma^{\rm in}$ — the same $4/3$ as the synchrotron power, as it must be. When instead $\gamma E_\gamma^{\rm in}\gtrsim m_e c^2$ the rest-frame photon is no longer soft, Step 2 fails, the cross section falls below $\sigma_T$ , and the $\gamma^2$ scaling is quenched: the Klein–Nishina regime. Why: naming the boundary matters as much as the result, since applying the Thomson formula past it badly overestimates the highest-energy losses.

Go deeper · Derivation C · 1 minFrom a power-law electron spectrum to a power-law photon spectrum

Derivation CFrom a power-law electron spectrum to a power-law photon spectrum

The purpose is to derive the observational cornerstone $\alpha=(p-1)/2$ — that an electron distribution $N(\gamma)\propto\gamma^{-p}$ radiates a flux $S_\nu\propto\nu^{-\alpha}$ — by treating each electron as emitting all its power at its critical frequency and integrating over the population. The same bookkeeping applies to inverse-Compton, so the two humps share a spectral index and a measured slope reports the electron index directly.

Step 1. Adopt the monochromatic (delta-function) approximation: each electron radiates its power $P\propto\gamma^2 U$ at $\nu_c\propto\gamma^2 B$ , so $\gamma\propto(\nu/B)^{1/2}$ . Why: the single-electron spectrum is broad but sharply peaked, so the delta approximation gets the ensemble slope exactly right — its width only smooths features, it does not change the power law.

Step 2. Write the emissivity as the population-weighted single-electron power in a frequency bin, $j_\nu\,\mathrm d\nu\propto N(\gamma)\,P\,\mathrm d\gamma$ evaluated where $\nu\propto\gamma^2$ ; the mapping gives $\mathrm d\gamma\propto\nu^{-1/2}\,\mathrm d\nu$ . Why: conservation of electrons across the change of variable from $\gamma$ to $\nu$ is the one step that carries the whole derivation.

Step 3. Collect the powers of $\gamma$ : $N(\gamma)\propto\gamma^{-p}$ , the power $P\propto\gamma^2$ , and the Jacobian $\mathrm d\gamma/\mathrm d\nu\propto\gamma^{-1}$ combine as $j_\nu\propto\gamma^{-(p-1)}$ . Substituting $\gamma\propto\nu^{1/2}$ gives $j_\nu\propto\nu^{-(p-1)/2}$ , so $\alpha=(p-1)/2$ — the shock value $p=2.5$ gives $\alpha=0.75$ , and $p=2$ gives $\alpha=0.5$ . Why: this is the bridge from microphysics to observation — a slope on a distant radio map directly reads out the electron index $p=2\alpha+1$ .

Limiting cases

The behaviour turns on three knobs: electron energy $\gamma$ , field strength, and whether the scattering is Thomson or Klein–Nishina.

Case	Result	Meaning
Nonrelativistic ( $\gamma\to1$ )	$\nu_c\to\nu_g=\dfrac{eB}{2\pi m_e c}$	The broad synchrotron continuum collapses to line emission at the gyrofrequency and its low harmonics — ordinary cyclotron radiation.
Thomson inverse-Compton	$\langle E_\gamma^{\rm out}\rangle=\tfrac{4}{3}\gamma^2 E_\gamma^{\rm in}$ , valid for $\gamma E_\gamma^{\rm in}\ll m_e c^2$	Cross section is the constant $\sigma_T$ ; energy gain scales cleanly as $\gamma^2$ and $P_{\rm IC}\propto\gamma^2 U_{\rm rad}$ .
Klein–Nishina	$\sigma_{\rm KN}\simeq\sigma_T\,\dfrac{3}{8x}\!\left(\ln 2x+\tfrac12\right),\ x=\dfrac{\gamma E_\gamma^{\rm in}}{m_e c^2}\gg1$	The cross section falls as $1/x$ ; the $\gamma^2$ scaling breaks and inverse-Compton losses are suppressed for the most energetic electrons.
Synchrotron self-absorption (low $\nu$ )	$S_\nu\propto\nu^{5/2}$	Below a turnover the source is optically thick to its own emission; the spectrum rises as $\nu^{5/2}$ , a slope independent of $p$ .
Radiation-dominated cooling	$\dfrac{L_{\rm IC}}{L_{\rm sync}}=\dfrac{U_{\rm rad}}{U_B}>1$	When the ambient photon field outweighs the magnetic field, inverse-Compton dominates the losses and can run away into the Compton catastrophe.

Common trapThe single-electron peak is not the ensemble peak, and the Thomson formula does not survive Klein–Nishina

Two traps sit close together. First, the critical frequency $\nu_c\propto\gamma^2 B$ is where one electron of a given $\gamma$ radiates most of its power — not where the observed spectrum of a whole source peaks. The ensemble spectrum is the convolution of $\nu_c$ over the whole electron distribution, and its shape — power law, self-absorbed turnover, cooling break — is set by the distribution, not by any single $\gamma$ . Second, the tidy Thomson results $P_{\rm IC}\propto\gamma^2 U_{\rm rad}$ and $\langle E_\gamma^{\rm out}\rangle=\tfrac{4}{3}\gamma^2 E_\gamma^{\rm in}$ hold only while the seed photon is soft in the electron rest frame, $\gamma E_\gamma^{\rm in}\ll m_e c^2$ . Once $x=\gamma E_\gamma^{\rm in}/m_e c^2\gtrsim1$ the cross section plunges below $\sigma_T$ and the Thomson scaling badly overestimates the gamma-ray output of the highest-energy electrons.

Worked examples

1 · A radio-lobe electron: what it emits and how long it lives (fully worked)

An electron with $\gamma=10^4$ sits in the $B=10\ \mu\mathrm{G}=10^{-9}\ \mathrm{T}$ field of a radio-galaxy lobe, pitch angle $\alpha=90^\circ$ . At what frequency does it radiate, and how long before it cools?

Step 1. Critical frequency. The gyrofrequency is $\nu_g=eB/(2\pi m_e c)\approx2.8\ \mathrm{MHz}$ per gauss, and $10\ \mu\mathrm{G}=10^{-5}$ G gives $\nu_g\approx28\ \mathrm{Hz}$ . Then $\nu_c=\tfrac{3}{2}\gamma^2\nu_g\sin\alpha=1.5\times10^8\times28\ \mathrm{Hz}$ . ν_c ≈ 4.2 GHz — in the centimetre radio band where such lobes are mapped.

Step 2. Magnetic energy density. $U_B=B^2/2\mu_0=(10^{-9})^2/(2\times4\pi\times10^{-7})\ \mathrm{J\,m^{-3}}$ . U_B ≈ 4.0×10⁻¹³ J m⁻³. For reference the CMB carries $U_{\rm CMB}=aT^4\approx4.2\times10^{-14}\ \mathrm{J\,m^{-3}}$ , a tenth of $U_B$ , so synchrotron dominates the cooling here (the lobe is above the $B_{\rm CMB}\approx3.2\ \mu$ G equivalent field).

Step 3. Cooling time from equation (6), using $U=U_B+U_{\rm CMB}\approx4.4\times10^{-13}\ \mathrm{J\,m^{-3}}$ : $t_{\rm cool}=3m_e c/(4\sigma_T\gamma U)$ . With $m_e c=2.73\times10^{-22}$ , $\sigma_T=6.65\times10^{-29}$ , $\gamma=10^4$ this gives $t_{\rm cool}\approx7\times10^{14}\ \mathrm{s}$ . t_cool ≈ 22 Myr. These electrons persist for tens of millions of years — long enough to travel far down a jet, which is why giant radio lobes glow so far from their nucleus.

2 · The same electron upscatters the microwave background (fully worked)

The same $\gamma=10^4$ electron scatters a CMB photon of energy $E_\gamma^{\rm in}\approx6\times10^{-4}$ eV. What photon comes out, is the Thomson limit valid, and what is the inverse-Compton-to-synchrotron luminosity ratio?

Step 1. Output energy from equation (5): $\langle E_\gamma^{\rm out}\rangle=\tfrac{4}{3}\gamma^2 E_\gamma^{\rm in}=\tfrac{4}{3}\times10^8\times6\times10^{-4}\ \mathrm{eV}$ . E_out ≈ 80 keV — a hard X-ray for Chandra or NuSTAR, the microwave photon promoted through eight decades in one scatter.

Step 2. Thomson check. The rest-frame photon energy is $\gamma E_\gamma^{\rm in}=10^4\times6\times10^{-4}=6\ \mathrm{eV}$ , versus $m_e c^2=511\ \mathrm{keV}$ , so $x\approx10^{-5}\ll1$ : the Thomson limit holds well and equation (5) applies with no correction.

Step 3. Luminosity ratio from equation (3): $L_{\rm IC}/L_{\rm sync}=U_{\rm CMB}/U_B=4.2\times10^{-14}/4.0\times10^{-13}$ . L_IC/L_sync ≈ 0.10. The lobe radiates about a tenth as much in inverse-Compton X-rays as in synchrotron radio — and inverting this, a measured X-ray-to-radio ratio yields $B$ , since $U_{\rm CMB}$ is known precisely — the cleanest magnetometry in extragalactic astronomy.

3 · Reading the electron index off a spectral slope

A jet knot shows a radio spectral index $\alpha=0.75$ that steepens to $\alpha=1.25$ above a break frequency. Interpret both numbers.

Step 1. Injection index. From equation (4), $p=2\alpha+1=\mathbf{2.5}$ — the standard mildly steep outcome of diffusive shock acceleration, so $N(\gamma)\propto\gamma^{-2.5}$ .

Step 2. The break. Cooling removes the highest- $\gamma$ electrons first, so above the frequency where $t_{\rm cool}$ equals the source age the photon index steepens by $\Delta\alpha=0.5$ : from $0.75$ to $\mathbf{1.25}$ , exactly the observed jump. The break frequency is a clock — with $B$ known, $\gamma_{\rm break}\propto1/(B^2 t_{\rm age})$ dates the population.

Physics of the portrait

Synchrotron radiation taught astronomers to read magnetic fields in the sky — and took forty years to go from a laboratory curiosity to the brightest objects in the universe.

Origins · 1 minsynchrotron radiation and how this was discovered

synchrotron radiation

from Schott 1912 to Schwinger 1949 to the Crab Nebula 1953 · nonthermal emission of relativistic electrons in cosmic magnetic fields

The radiation of a charge on a circular orbit was worked out classically by George Adolphus Schott in 1912, in a treatise so far ahead of its time that it sat almost unused for decades. The problem returned with the accelerators of the 1940s: electrons whirling in the new synchrotrons lost energy alarmingly fast to radiation, and in 1947 a blue-white glow was seen streaming from the beam tube of a General Electric machine — which gave the effect its name. Julian Schwinger, better known for quantum electrodynamics, put the theory on a rigorous footing in 1949, deriving the full spectral distribution and the frequency $\nu_c\propto\gamma^2 B$ that organizes this whole article. The astrophysical leap came at once: in 1950 Alfvén and Herlofson proposed that the cosmic radio sources were relativistic electrons radiating in magnetic fields, and in 1953 Iosif Shklovsky argued that even the polarized, featureless optical continuum of the Crab Nebula was synchrotron — confirmed when the light was found polarized just as the theory demanded. Inverse-Compton entered in parallel, proposed by Feenberg and Primakoff in 1948 as an energy-loss channel for cosmic-ray electrons and made universal in 1965, when the microwave background gave every electron in the cosmos a guaranteed bath of seed photons.

History

1912

— Schott computes the classical radiation of a charge on a circular orbit

1949

— Schwinger derives the full synchrotron spectrum and its ν_c ∝ γ² B scaling

1950

— Alfvén and Herlofson attribute cosmic radio sources to synchrotron emission

1953

— Shklovsky explains the Crab Nebula's polarized optical continuum as synchrotron

1965

— Discovery of the microwave background gives inverse-Compton a universal seed field

Going further

Texts. Rybicki & Lightman, Radiative Processes in Astrophysics, chapters 6 (synchrotron) and 7 (Compton) — the definitive derivation of every formula here · Longair, High Energy Astrophysics, 3rd ed., chapters 8–9 · Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., on active galaxies · Ghisellini, Radiative Processes in High Energy Astrophysics · Dodelson, Modern Cosmology, for the Compton/CMB machinery.

Provenance. Every equation is a standard result checkable against Rybicki & Lightman: (1) the power coefficient $\tfrac{4}{3}$ from the relativistic Larmor formula with $\sigma_T=6.652\times10^{-29}\ \mathrm{m^2}$ ; (2) $\alpha=(p-1)/2$ from integrating $N(\gamma)\propto\gamma^{-p}$ over $\nu\propto\gamma^2 B$ ; (3) the $\gamma^2$ inverse-Compton gain from two Doppler boosts. The worked numbers — $\nu_c\approx4.2$ GHz, $t_{\rm cool}\approx22$ Myr, $E_{\rm out}\approx80$ keV, $L_{\rm IC}/L_{\rm sync}\approx0.10$ for a $\gamma=10^4$ electron in a $10\ \mu$ G field bathed in the CMB — are recomputed in verification; the $\sim10^{12}$ K Compton ceiling and $B_{\rm CMB}\approx3.2\ \mu$ G equivalent field are the standard AGN-literature values.

Prerequisites. Special relativity and Lorentz transformations · Classical electromagnetism and the Larmor formula · Thomson and Compton scattering · Relativistic jets and their launching. Leads to. Blazars and relativistic beaming · Active galactic nuclei: the unified model · Quasars and quasar evolution.

Ask the tutor: Ask me to walk through fitting a real blazar's double-humped spectrum — extracting the field, Doppler factor, and electron index from the two peaks — or to derive any formula here from the relativistic Larmor law.

Why are synchrotron and inverse-Compton treated as one topic instead of two?

Because they are the same physics in different clothes. In both, a relativistic electron Thomson-scatters a field: the static magnetic field acts through its Lorentz force exactly as a photon field of energy density $U_B=B^2/2\mu_0$ would, and a real photon bath acts through $U_{\rm rad}$ . The relativistic Larmor calculation therefore yields the identical power $P=\tfrac{4}{3}\sigma_T c\,\gamma^2\beta^2 U$ in both cases, with only $U$ changing. This is not a loose analogy but an exact correspondence, whose useful consequence is that dividing the two luminosities gives $U_{\rm rad}/U_B$ — a magnetic-field measurement with no model-dependent factors.

What sets the characteristic synchrotron frequency, and why the γ² and not just γ?

Two independent relativistic effects each contribute a factor of $\gamma$ . First, the emission is beamed into a forward cone of half-angle $\sim1/\gamma$ , so the observer is illuminated for only $\sim1/\gamma$ of each orbit. Second, during that exposure the electron chases its own radiation at nearly $c$ , compressing the pulse in arrival time by another factor $\sim1/\gamma$ — the same time-compression that shortens the received period of an approaching source. A pulse shortened by $1/\gamma^2$ contains Fourier components up to $\gamma^2$ times the gyrofrequency, so $\nu_c\propto\gamma^2 B$ — the $\gamma^2$ genuinely the product of beaming and time-compression.

How does inverse-Compton let a microwave photon become an X-ray?

Through two Doppler boosts. Transform into the electron's rest frame and the incoming soft photon is blueshifted by a factor of order $\gamma$ . In that frame the electron is at rest, so the scattering is ordinary Thomson scattering — nearly elastic, changing the photon's direction but little of its energy. Transform back to the laboratory and the outgoing photon is blueshifted again by another factor of order $\gamma$ . The two boosts multiply, so the photon emerges with about $\gamma^2$ times its original energy. For $\gamma=10^4$ that is $10^8$ , turning a $6\times10^{-4}$ eV microwave photon into an 80 keV X-ray — the electron stamps its own energy onto every photon it touches.

Why does a power-law electron spectrum give a power-law photon spectrum?

Because the mapping between electron energy and emitted frequency is itself a power law: each electron radiates near $\nu\propto\gamma^2$ . Feed in $N(\gamma)\propto\gamma^{-p}$ and change variables from $\gamma$ to $\nu$ ; the electron count, the $\gamma^2$ power, and the Jacobian combine to a single power of $\nu$ , giving $S_\nu\propto\nu^{-(p-1)/2}$ . In the optically-thin, uncooled regime there are no other scales, so nothing turns the power law into anything else — which is why it is the telltale sign of nonthermal emission, its slope reporting $p=2\alpha+1$ .

What is the Compton catastrophe and why does it cap brightness temperature?

If a synchrotron source is made very compact and bright, the density of its own photons rises, and those photons become seeds for inverse-Compton scattering by the very electrons that made them — synchrotron-self-Compton. The internal $U_{\rm rad}$ climbs steeply with brightness temperature; once it approaches $U_B$ the inverse-Compton losses match the synchrotron losses, and pushing further spawns a second scattering generation that drains the electrons catastrophically fast. The runaway sets in near $10^{12}$ K. Real blazar cores appear hotter, which is a signature rather than a violation: their emission is Doppler-boosted toward us by an approaching jet, so the intrinsic temperature stays below the limit while the apparent one exceeds it.

Why do the highest-energy electrons die first, and what does that leave in the spectrum?

The radiated power grows as $\gamma^2$ while the energy grows only as $\gamma$ , so the cooling time $t_{\rm cool}=E/P\propto1/\gamma$ shrinks with energy and the most energetic electrons radiate themselves away soonest. In a source of finite age this depletes the distribution above the $\gamma$ whose cooling time equals the source age, steepening the photon spectrum by $\Delta\alpha=0.5$ . The frequency where this break appears is a clock: with a field estimate it dates the population, which is how radio astronomers age jet knots and radio-galaxy lobes.