Synchrotron Radiation

Point a radio telescope at almost any violent place in the universe — the shell of an exploded star, the light-year-long jet spat out by a black hole, the tangled magnetic web of our own Galaxy — and the glow that comes back is almost never thermal. It does not carry the smooth Planck curve of a hot body; it rises or falls as a clean power law across decade after decade of frequency, and it arrives strongly polarized. That radiation is synchrotron emission: the light of relativistic electrons spiralling in magnetic fields, and it is the single most important nonthermal process in high-energy astrophysics. It is the ultrarelativistic cousin of cyclotron radiation. A slow electron in a field $B$ hums at one gyration frequency $\nu_B=eB/(2\pi m_e c)$ ; crank the electron up to a Lorentz factor $\gamma\gg1$ and relativistic beaming crushes its emission into a forward searchlight of half-angle $1/\gamma$ and smears the spectrum into a broad continuum reaching from radio to X-rays. Because the radiated power scales as $m^{-4}$ at fixed energy, electrons — $1836$ times lighter than protons — do essentially all the shining. The magic of the process is that the shape of the light encodes the shape of the particle population: fold the single-electron spectrum over a power-law electron distribution $N(\gamma)\propto\gamma^{-p}$ and the sky hands you back $I_\nu\propto\nu^{-(p-1)/2}$ . Measure one number, the spectral index, and you have measured how the cosmos accelerates its particles. This article builds that machinery from the geometry of a single beamed pulse up to the four diagnostic fingerprints — power-law slope, polarization, self-absorption turnover, and cooling break — that let us read magnetic fields and particle energies off objects billions of light-years away.

Key takeaways

Synchrotron radiation is emitted by relativistic electrons gyrating in magnetic fields; it is the relativistic limit of the cyclotron problem and produces a beamed, broadband, nonthermal continuum from radio to X-rays.
A single electron of Lorentz factor $\gamma$ beams into a cone of half-angle $\sim1/\gamma$ , so the observer sees a short pulse whose spectrum peaks near the critical frequency $\nu_c\propto\gamma^2 B$ .
A power-law electron distribution $N(\gamma)\,\mathrm{d}\gamma\propto\gamma^{-p}\,\mathrm{d}\gamma$ maps to a power-law spectrum $I_\nu\propto\nu^{-\alpha}$ with spectral index $\alpha=(p-1)/2$ — the diagnostic signature that ties emission to particle acceleration.
The emission is intrinsically and strongly linearly polarized, up to $\Pi=(p+1)/(p+7/3)\approx70$ – $75\%$ for a uniform field, tracing the projected magnetic geometry.
At low frequency the source turns optically thick to synchrotron self-absorption, giving a $I_\nu\propto\nu^{5/2}$ rise independent of $p$ ; at high frequency radiative cooling steepens the slope by $\Delta\alpha=1/2$ above a break frequency $\nu_b$ .
It powers supernova remnants, radio galaxies and their jets, pulsar wind nebulae, and the Galactic radio foreground that must be scrubbed from the cosmic microwave background — the workhorse of nonthermal astrophysics.
Key idea: a relativistic electron beams its light into a cone of half-angle $1/\gamma$ , so the observer catches a pulse of duration $\sim1/(\gamma^2\nu_B)$ whose Fourier transform peaks near $\nu_c\simeq\tfrac32\gamma^2\nu_B\sin\alpha$ ; convolve that single-particle spectrum with $N(\gamma)\propto\gamma^{-p}$ and you get the emergent power law $I_\nu\propto\nu^{-(p-1)/2}$ .

The core insight

A relativistic electron beams its radiation into a cone of half-angle $1/\gamma$ ; the observer sees a brief pulse of duration $\sim1/(\gamma^2\nu_B)$ whose Fourier transform is a broad spectrum peaking near $\nu_c\simeq\tfrac32\gamma^2\nu_B\sin\alpha$ , and integrating that single-particle spectrum over a power-law electron distribution $N(\gamma)\propto\gamma^{-p}$ yields the emergent power law $I_\nu\propto\nu^{-(p-1)/2}$ .

Where this lives today

Synchrotron radiation is the currency of an entire branch of observational astrophysics, and the current generation of instruments is built around reading it. The Low-Frequency Array (LOFAR) and the forthcoming Square Kilometre Array map the steep-spectrum, aged synchrotron plasma of radio galaxies and cluster halos at tens to hundreds of MHz, precisely where cooling breaks and self-absorption turnovers live, letting astronomers date jet outbursts and weigh the magnetic and particle energy of the largest structures in the universe. The Event Horizon Telescope images the innermost self-absorbed synchrotron emission around the black holes in M87 and the Galactic centre, where the $\nu^{5/2}$ physics of a compact core sets the size and brightness of the ring. Polarization is having its own renaissance: the same physics that makes synchrotron light up to $75\%$ polarized makes the diffuse Galactic synchrotron foreground the chief obstacle for experiments hunting the primordial $B$ -mode polarization of the cosmic microwave background, so missions like the Simons Observatory and LiteBIRD must model and subtract it exquisitely — a direct line from a single spiralling electron to the physics of inflation. And in the X-ray band, Chandra's thin synchrotron rims on supernova remnants and IXPE's X-ray polarimetry of those same rims reveal freshly accelerated, rapidly cooling PeV-scale electrons, pinning down where and how galactic cosmic rays are born. Meanwhile the process runs in reverse in the laboratory: synchrotron light sources and X-ray free-electron lasers deliberately harness the same emission to probe matter, closing the loop between the cosmic and the engineered.

The physical picture

Picture one electron riding a tight helix around a magnetic field line. In its own instantaneous rest frame it does nothing exotic — it feels a transverse acceleration and radiates the gentle doughnut pattern of an oscillating dipole, sending light out mostly sideways. Now boost to the lab frame at $\gamma\gg1$ . Relativistic aberration takes that lazy dipole doughnut and crushes it forward into a needle-thin cone of half-angle $\sim1/\gamma$ aimed along the velocity. The consequence is dramatic: you, the distant observer, are dark for almost the entire orbit and receive light only during the sliver of time when that searchlight sweeps directly across your line of sight. Synchrotron light reaches you not as a steady glow but as a train of ferociously brief flashes, one per gyration, each lasting a tiny fraction of the orbital period.

Figure

Relativistic beaming: an electron of Lorentz factor γ spirals around B at pitch angle α and radiates into a forward cone of half-angle 1/γ. Only during the swept arc Δθ ≈ 2/γ does the observer see light. For γ = 10⁴ in a B = 10 μG field the peak lands at ν_c ≈ 4.2 GHz, deep in the radio

Why does one electron light up the whole radio-to-X-ray band? Two compressions stack. First, the searchlight only illuminates you over an orbital arc $\Delta\theta\sim2/\gamma$ , so the emission is already brief. Second — and this is the deep one — the electron is chasing its own pulse at nearly $c$ , so the leading and trailing edges of the flash arrive almost on top of each other, squeezing the observed duration by a further factor $(1-\beta)\approx1/(2\gamma^2)$ . A pulse shortened to $\sim1/(\gamma^2\nu_B)$ has, by the Fourier uncertainty principle, a spectrum broadened to $\nu_c\sim\gamma^2\nu_B$ . That single $\gamma^2$ is the engine of the whole subject: it turns a $\sim$ MHz gyrofrequency in a microgauss field into GHz radio waves and, for the most energetic electrons, into hard X-rays. Stack many electrons with a spread of energies and each contributes a bump near its own $\nu_c\propto\gamma^2$ ; weight those bumps by $\gamma^{-p}$ and they blur into a featureless power law whose slope is inherited directly from the particle distribution. The second figure shows the composite spectrum every real source displays — a low-frequency self-absorbed rise, an optically-thin power law, and a cooling break where the highest-energy electrons have already burned out.

Figure

The broadband synchrotron spectrum on log-log axes: a self-absorbed ν^(5/2) rise (independent of p) climbs to a turnover, then an optically-thin power law of slope −α with α = (p−1)/2 = 0.5 for p = 2, which steepens by Δα = 1/2 to −1.0 above the cooling break ν_b

Polarization is the fourth fingerprint, and it is pure geometry. The electron's acceleration is perpendicular to $\mathbf{B}$ , so the electric vector of the light it radiates remembers the field direction. A well-ordered field therefore yields strongly, linearly polarized emission whose plane maps the projected field; a tangled field, or Faraday rotation smearing the polarization angle along the line of sight, washes that signal out. Where synchrotron polarization survives, it is the most direct probe of cosmic magnetic geometry we possess.

The core equations

Start with the rhythm of the motion. An electron of Lorentz factor $\gamma$ in a field $B$ gyrates at the relativistic gyrofrequency — the nonrelativistic cyclotron rate divided by $\gamma$ , because the relativistic mass is $\gamma m_e$ .

\nu_B=\dfrac{eB}{2\pi\gamma m_e c}\qquad(\text{relativistic gyrofrequency}) \qquad (1)

where: $e$ — electron charge magnitude; $B$ — magnetic flux density; $\gamma$ — Lorentz factor; $m_e$ — electron rest mass; $c$ — speed of light. Numerically the nonrelativistic rate is $\nu_{B,0}=eB/(2\pi m_e c)\approx2.8\,\text{MHz}\,(B/1\,\text{G})$ , so in a $10\,\mu$ G interstellar field it is a mere $\sim28$ Hz — the beaming boost, not the gyration, is what makes the radio light. The observed spectrum peaks not at $\nu_B$ but at the critical frequency, boosted by the all-important $\gamma^2$ from beaming and pulse compression and modulated by the pitch angle $\alpha$ between the velocity and the field.

\nu_c=\dfrac{3}{2}\,\gamma^2\,\nu_{B,0}\,\sin\alpha=\dfrac{3\gamma^2 eB\sin\alpha}{4\pi m_e c} \qquad (2)

where: $\nu_{B,0}$ — nonrelativistic gyrofrequency $eB/(2\pi m_e c)$ ; $\alpha$ — pitch angle (the angle between $\mathbf{v}$ and $\mathbf{B}$ ); $\nu_c$ — critical frequency near which the single-electron spectrum peaks. Reading it: $\nu_c\approx4.2\,\text{MHz}\,(B/1\,\text{G})\,\gamma^2\sin\alpha$ , so $\gamma=10^4$ electrons in a $10\,\mu$ G field radiate their peak at a few GHz — squarely in the radio. How much energy does one electron pour out? Averaging over pitch angle, the total radiated power takes a compact form in terms of the Thomson cross section and the magnetic energy density.

P=\dfrac{4}{3}\,\sigma_T\,c\,\beta^2\,\gamma^2\,U_B,\qquad U_B=\dfrac{B^2}{8\pi} \qquad (3)

where: $P$ — total radiated power of one electron; $\sigma_T=6.65\times10^{-25}\,\mathrm{cm^2}$ — Thomson cross section; $U_B$ — magnetic energy density; $\beta=v/c$ . The $\gamma^2$ here is why energetic electrons cool fast, and the identical functional form to inverse-Compton power (with $U_B\to U_{\rm rad}$ ) is why the two processes are twins. Now the step that makes synchrotron a diagnostic tool: convolve the single-electron emissivity with the electron energy distribution. For a power law $N(\gamma)\propto\gamma^{-p}$ the integral collapses to another power law.

j_\nu\propto\int N(\gamma)\,P_\nu(\gamma)\,\mathrm{d}\gamma\;\Rightarrow\;j_\nu\propto\nu^{-(p-1)/2}\quad\text{for}\;N(\gamma)\propto\gamma^{-p} \qquad (4)

where: $j_\nu$ — emission coefficient (power per unit volume, frequency, solid angle); $N(\gamma)$ — electron number density per unit $\gamma$ ; $p$ — electron power-law index; $P_\nu$ — single-electron spectral power. The spectrum literally inherits the slope of the particle population — measure the light, learn the particles. In an optically thin source the emergent intensity follows the emissivity, defining the observed spectral index.

I_\nu\propto\nu^{-\alpha},\qquad\alpha=\dfrac{p-1}{2} \qquad (5)

where: $\alpha$ — spectral index (the power-law slope, NOT the pitch angle, which shares the symbol); $I_\nu$ — emergent specific intensity. Diffusive shock acceleration predicts $p\approx2$ , hence $\alpha\approx0.5$ — exactly the slope seen across supernova remnants and radio jets. At low frequency the source absorbs its own radiation. The source function then approaches a Rayleigh-Jeans-like form at the effective temperature of the radiating electrons, giving a slope that is completely independent of $p$ .

S_\nu=\dfrac{j_\nu}{\alpha_\nu}\propto\nu^{5/2},\qquad I_\nu\xrightarrow{\;\tau_\nu\gg1\;}S_\nu \qquad (6)

where: $S_\nu$ — source function $j_\nu/\alpha_\nu$ ; $\alpha_\nu$ — synchrotron self-absorption coefficient; $\tau_\nu$ — optical depth to self-absorption. The $\nu^{5/2}$ rise and its turnover locate the self-absorption frequency and, with it, the source's brightness temperature and magnetic field. Ordered fields imprint linear polarization, whose optically-thin ceiling depends only on the spectral slope.

\Pi=\dfrac{p+1}{p+\tfrac{7}{3}}\qquad(\text{optically thin, uniform }B) \qquad (7)

where: $\Pi$ — fractional linear polarization; $p$ — electron power-law index. For $p=2$ this is $69\%$ , for $p=3$ it is $75\%$ — far above anything thermal emission can produce. Finally, energetic electrons burn out. The cooling time is the energy divided by the radiated power, and it falls with both particle energy and field strength.

t_{\rm cool}=\dfrac{E}{P}=\dfrac{\gamma m_e c^2}{\tfrac{4}{3}\sigma_T c\,\gamma^2 U_B}\propto\dfrac{1}{\gamma B^2} \qquad (8)

where: $t_{\rm cool}$ — synchrotron cooling timescale; $E=\gamma m_e c^2$ — electron energy. Since the electrons radiating at higher $\nu$ have higher $\gamma$ , the top of the spectrum ages first, producing a moving cooling break.

\nu_b\;(\text{cooling break}):\quad\alpha\to\alpha+\tfrac{1}{2}\;\;\text{for}\;\nu>\nu_b \qquad (9)

where: $\nu_b$ — break frequency where $t_{\rm cool}(\gamma)$ equals the source age. Above $\nu_b$ the spectrum steepens by exactly one half — an on-sky clock for the age of the plasma.

Derivation ABeaming and the critical frequency

The goal is to show, from pure kinematics and with no appeal to the full Bessel-function theory, that the peak emission frequency of a single relativistic electron scales as $\nu_c\propto\gamma^2 B$ — the central scaling of the whole subject. We track the arrival times of the leading and trailing edges of the beamed pulse an observer receives, combine the width of the emission cone with the retarded-time compression from the electron chasing its own light, and read off the pulse duration whose reciprocal is the peak frequency.

Step 1. The observer is illuminated only while the velocity vector sweeps through the emission cone, of angular full-width $\Delta\theta\approx2/\gamma$ (half-angle $1/\gamma$ on each side). Why: relativistic aberration compresses the rest-frame dipole into this forward cone, so light reaches a fixed observer only across this small arc of the orbit.

Step 2. The electron takes an emission interval $\Delta t_e=\Delta\theta/\omega_B=2/(\gamma\,\omega_B)$ to traverse that arc, where $\omega_B=eB\sin\alpha/(\gamma m_e c)$ is the relativistic angular gyrofrequency projected onto the plane perpendicular to $\mathbf{B}$ . Why: this is just arc-angle over angular speed, converting the geometric cone width into a rest-frame duration.

Step 3. The observed pulse is far shorter than $\Delta t_e$ because the electron nearly keeps pace with its own light: $\Delta t_o=\Delta t_e(1-\beta\cos\theta)\approx\Delta t_e(1-\beta)\approx\Delta t_e/(2\gamma^2)$ . Why: the trailing edge is emitted from a point that has moved almost a full pulse-length closer to the observer, so the two edges arrive nearly together — this is the Doppler pile-up that supplies the second factor of $\gamma$ .

Step 4. Assembling the two compressions gives $\Delta t_o\approx1/(\gamma^3\omega_B)=m_e c/(\gamma^2 eB\sin\alpha)$ , so the peak frequency is $\nu_c\sim1/(2\pi\Delta t_o)\sim\gamma^2 eB\sin\alpha/(2\pi m_e c)$ . Why: a pulse of width $\Delta t_o$ has a Fourier spectrum extending to $\sim1/\Delta t_o$ ; the exact theory supplies the leading factor $3/2$ , recovering equation (2). *Verified: the scaling $\nu_c\propto\gamma^2 B$ and the $3/2$ prefactor match Rybicki & Lightman §6.2 and Carroll & Ostlie §4.3; numerically $\nu_c\approx4.2\,\text{MHz}\,(B/\text{G})\,\gamma^2\sin\alpha$ reproduces the anchor $\nu_c\approx4.2$ GHz for $\gamma=10^4$ , $B=10\,\mu$ G.*

Go deeper · Derivation B · 1 minPower-law electrons make a power-law spectrum

Derivation BPower-law electrons make a power-law spectrum

The aim here is the result that turns synchrotron radiation from a curiosity into astrophysics' finest particle-acceleration diagnostic: a power-law electron distribution $N(\gamma)\propto\gamma^{-p}$ radiates a power-law spectrum $I_\nu\propto\nu^{-(p-1)/2}$ , so the observed spectral index $\alpha$ measures the particle index $p$ directly. We approximate each electron as radiating all its power at its own critical frequency (the delta-function approximation, exact enough for the exponent), insert the power-law distribution, change variables from $\gamma$ to $\nu$ , and collect the powers of $\nu$ and $B$ .

Step 1. Each electron radiates total power $P\propto\gamma^2 B^2$ concentrated near $\nu_c=C\gamma^2 B$ with $C=\tfrac32(e\sin\alpha)/(4\pi m_e c/\ldots)$ a constant, so approximate $P_\nu\approx P\,\delta(\nu-\nu_c)$ . Why: the true single-electron spectrum is broad but sharply peaked near $\nu_c$ ; replacing it by a delta at $\nu_c$ preserves the power-law exponent while making the integral trivial.

Step 2. Insert $N(\gamma)=N_0\gamma^{-p}$ and integrate: $j_\nu\propto\int_0^\infty\gamma^{-p}\,\gamma^2 B^2\,\delta(\nu-C\gamma^2 B)\,\mathrm{d}\gamma$ . Why: the emissivity is the electron distribution weighted by each electron's spectral power, summed over all energies.

Step 3. The delta selects $\gamma=(\nu/CB)^{1/2}$ , and its Jacobian contributes $\mathrm{d}\gamma=\tfrac12(CB)^{-1/2}\nu^{-1/2}\,\mathrm{d}\nu$ . Why: evaluating a delta function $\delta(\nu-C\gamma^2 B)$ against $\mathrm{d}\gamma$ divides by $|\mathrm{d}(C\gamma^2B)/\mathrm{d}\gamma|=2C\gamma B$ , which supplies the extra $\nu^{-1/2}$ .

Step 4. Substituting $\gamma\propto(\nu/B)^{1/2}$ everywhere and collecting gives $j_\nu\propto B^2\cdot\nu^{(2-p)/2}B^{-(2-p)/2}\cdot\nu^{-1/2}B^{-1/2}=B^{(p+1)/2}\,\nu^{-(p-1)/2}$ . Why: gathering the $\gamma^{2-p}$ factor, the $B^2$ from the power, and the Jacobian yields both the field scaling (used for equipartition estimates) and the frequency slope.

Step 5. Hence $\alpha=(p-1)/2$ : for $p=2$ , $\alpha=0.5$ ; for $p=3$ , $\alpha=1$ — the observed radio slopes. Why: this closes the loop, mapping the microscopic acceleration index onto the macroscopic, directly measured spectral index. *Verified: reproduces Rybicki & Lightman Eq. (6.36) and Longair Ch. 8; the exponent algebra gives $p=2\Rightarrow\alpha=0.5$ and $p=3\Rightarrow\alpha=1.0$ as checked in the worked examples.*

Go deeper · Derivation C · 1 minSynchrotron cooling and the spectral break

Derivation CSynchrotron cooling and the spectral break

The purpose of this derivation is to explain why aged synchrotron sources show a spectral break above which the slope steepens by exactly $\Delta\alpha=1/2$ , and why that break marches to lower frequency as the source ages — the standard clock used to date radio lobes and remnants. We balance the energy-dependent cooling time against the source age to find the electron energy above which the population is depleted, translate the depletion into a steepened electron index under continuous injection, and map that back onto the photon spectrum.

Step 1. The cooling time $t_{\rm cool}(\gamma)=\gamma m_e c^2/(\tfrac43\sigma_T c\,\gamma^2 U_B)\propto\gamma^{-1}$ falls with energy, so the most energetic electrons burn out first. Why: radiated power grows as $\gamma^2$ while energy grows only as $\gamma$ , so the ratio — the survival time — shrinks as $1/\gamma$ .

Step 2. Setting $t_{\rm cool}(\gamma_b)=t_{\rm age}$ defines a break Lorentz factor $\gamma_b\propto(B^2 t_{\rm age})^{-1}$ above which electrons have already radiated their energy away. Why: an electron survives to the present only if its cooling time exceeds the source age; the equality marks the boundary.

Step 3. Under continuous injection with losses, the steady-state distribution above $\gamma_b$ steepens by one power: $N(\gamma)\propto\gamma^{-(p+1)}$ for $\gamma>\gamma_b$ . Why: in steady state the cooling flux in energy space piles electrons up just below $\gamma_b$ and thins them above it, adding exactly one to the exponent.

Step 4. Feeding the steepened index through $\alpha=(p-1)/2$ gives $\alpha=[(p+1)-1]/2=p/2=\alpha_{\rm inj}+\tfrac12$ , and $\nu_b\propto\gamma_b^2 B\propto B^{-3}t_{\rm age}^{-2}$ . Why: the half-integer jump is model-independent for continuous injection, and the $B^{-3}t^{-2}$ scaling lets an observed break frequency date the source once the field is estimated. *Verified: consistent with the Kardashev-Pacholczyk continuous-injection aging framework and Longair Ch. 8; the change $\Delta\alpha=1/2$ and $\nu_b\propto B^{-3}t^{-2}$ reproduce the aged-jet example below.*

Limiting cases

Synchrotron behaviour is set by three knobs: the electron Lorentz factor $\gamma$ , the field strength $B$ , and the optical depth $\tau_\nu$ . The regimes:

Case	Result	Meaning
Nonrelativistic (cyclotron)	$\gamma\to1,\ \nu\to\nu_B=\dfrac{eB}{2\pi m_e c}$	Emission collapses to the single gyrofrequency and its low harmonics; no beaming, no continuum. The gyroresonance limit.
Mildly relativistic (gyrosynchrotron)	$\gamma\sim\text{few},\ \nu\sim(\text{few to}\sim\gamma^2)\,\nu_B$	Discrete harmonics broaden and blend; the regime of solar flares and stellar coronae, transitional to true synchrotron.
Ultrarelativistic (synchrotron)	$\gamma\gg1,\ \nu_c=\dfrac{3}{2}\gamma^2\nu_B\sin\alpha$	Beaming makes a broad continuum peaking at $\nu_c$ ; a power-law population gives $I_\nu\propto\nu^{-(p-1)/2}$ . The astrophysical regime.
Optically thick (self-absorbed)	$\tau_\nu\gg1:\ I_\nu\propto\nu^{5/2}$	Below the turnover the intensity follows $S_\nu\propto\nu^{5/2}$ , independent of $p$ — the rising low-frequency spectrum of compact radio cores.
Cooled (above the break)	$\nu>\nu_b:\ \alpha\to\dfrac{p}{2}\ (\Delta\alpha=\tfrac12)$	Radiative losses deplete high- $\gamma$ electrons, steepening the spectrum above the cooling break $\nu_b$ .

Common trapTwo symbols named α, and a slope that is not universal

The single most common slip in this subject is that $\alpha$ means two different things: the spectral index (the slope, $I_\nu\propto\nu^{-\alpha}$ ) and the pitch angle (the angle between $\mathbf{v}$ and $\mathbf{B}$ ). They are unrelated; keep them straight by context. A second trap: the single-electron spectrum does not peak at $\nu_c$ — the emission function $F(x)$ peaks near $x=\nu/\nu_c\approx0.29$ , so the actual maximum sits below the nominal critical frequency. Third, and most dangerous for interpreting data, the headline relation $\alpha=(p-1)/2$ holds only in the optically thin regime. Below the self-absorption turnover the spectrum is $I_\nu\propto\nu^{5/2}$ regardless of $p$ , because self-absorption drives the source toward a Rayleigh-Jeans-like effective temperature; applying $\alpha=(p-1)/2$ there will give you a nonsensical particle index. Always establish which side of the turnover you are on before inverting a slope into a $p$ .

Worked examples

1 · From spectral index to particle index (confidence builder)

A supernova remnant shows an optically-thin radio spectral index $\alpha=0.5$ . A radio galaxy's aged lobe shows $\alpha=1.0$ . What electron power-law index $p$ does each imply, and what does each tell you about the electrons?

Step 1. Invert equation (5), $\alpha=(p-1)/2\Rightarrow p=2\alpha+1$ . For the remnant, $p=2(0.5)+1=\mathbf{2.0}$ . This is the canonical value predicted by diffusive shock acceleration at a strong, non-relativistic shock — the light is telling you the shock is doing textbook Fermi-I acceleration.

Step 2. For the lobe, $p=2(1.0)+1=\mathbf{3.0}$ . A steep spectrum like this signals either intrinsically steep injection or, far more commonly, a strongly aged population whose high-energy end has cooled — the two are distinguished by looking for a break rather than a single slope.

Step 3. Sanity check the polarization ceiling for each with equation (7): for $p=2$ , $\Pi=(2+1)/(2+7/3)=3/4.33=\mathbf{69\%}$ ; for $p=3$ , $\Pi=4/5.33=\mathbf{75\%}$ . A well-ordered field could polarize these sources up to $\sim70$ – $75\%$ ; observing far less flags a tangled field or Faraday depolarization along the line of sight.

2 · Peak frequency and cooling time of a jet electron (fully worked)

An electron of Lorentz factor $\gamma=5.0\times10^3$ spirals with pitch angle $\alpha=90^\circ$ in the $B=10\,\mu\text{G}=1.0\times10^{-5}\,\text{G}$ field of an extragalactic jet. (a) At what frequency does it radiate its peak? (b) What is its energy? (c) How long can it keep shining before it cools? Use $\sigma_T=6.65\times10^{-25}\,\text{cm}^2$ , $m_e c^2=8.19\times10^{-7}\,\text{erg}$ , $c=3.0\times10^{10}\,\text{cm/s}$ .

Step 1. Peak frequency from equation (2) in the handy numerical form $\nu_c\approx4.2\,\text{MHz}\,(B/1\,\text{G})\,\gamma^2\sin\alpha$ . With $\sin90^\circ=1$ : $\nu_c=4.2\times10^6\times(1.0\times10^{-5})\times(5.0\times10^3)^2=4.2\times10^6\times10^{-5}\times2.5\times10^7$ . This is $\nu_c=\mathbf{1.05\times10^9\,\text{Hz}\approx1.0\ \text{GHz}}$ — a standard radio observing band.

Step 2. Energy: $E=\gamma m_e c^2=5.0\times10^3\times8.19\times10^{-7}\,\text{erg}=4.1\times10^{-3}\,\text{erg}$ . Converting, $4.1\times10^{-3}\,\text{erg}/(1.6\times10^{-6}\,\text{erg/TeV})\approx\mathbf{2.6\ \text{GeV}}$ . These are mildly-relativistic-by-jet-standards electrons; the X-ray-emitting ones sit at $\gamma\sim10^7$ – $10^8$ .

Step 3. Magnetic energy density: $U_B=B^2/8\pi=(1.0\times10^{-5})^2/(8\pi)=1.0\times10^{-10}/25.1=\mathbf{4.0\times10^{-12}\,\text{erg/cm}^3}$ .

Step 4. Cooling time from equation (8), written as $t_{\rm cool}=m_e c^2/(\tfrac43\sigma_T c\,\gamma U_B)$ . Denominator: $\tfrac43\times6.65\times10^{-25}\times3.0\times10^{10}\times5.0\times10^3\times4.0\times10^{-12}=5.3\times10^{-22}$ . So $t_{\rm cool}=8.19\times10^{-7}/5.3\times10^{-22}=1.55\times10^{15}\,\text{s}$ .

Step 5. Convert: $1.55\times10^{15}\,\text{s}/(3.16\times10^{7}\,\text{s/yr})\approx\mathbf{4.9\times10^{7}\ \text{yr}\approx50\ \text{Myr}}$ . A GHz-emitting electron in a microgauss jet field lives $\sim50$ Myr — comparable to the dynamical age of large radio lobes, which is exactly why cooling breaks appear right in the observable radio band and can be used to date the plasma. Note the scaling: an X-ray-emitting electron at $\gamma\sim10^8$ in the same field cools $\sim10^4$ times faster, in only thousands of years — so X-ray synchrotron traces acceleration happening essentially now.

Physics of the portrait

Synchrotron radiation is named for the machine that first made it visible — the electron synchrotron — but its theory reaches back to the classical electrodynamics of accelerated charges, and its astrophysical meaning was decoded by a separate cast of Soviet and Western theorists in the 1950s.

Origins · 1 minJulian Schwinger and how this was discovered

Julian Schwinger

1918–1994 · the classical theory of radiation from relativistic electrons

Julian Schwinger is remembered first for quantum electrodynamics — he shared the 1965 Nobel Prize with Feynman and Tomonaga for taming its infinities — but he also wrote the paper that put synchrotron radiation on a rigorous footing. In 1949, in "On the Classical Radiation of Accelerated Electrons" (Physical Review 75, 1912), he worked out the full angular and spectral distribution of the light from an electron on a curved relativistic path: the beaming into a $1/\gamma$ cone, the broadband spectrum peaking near the critical frequency, the total power scaling as $\gamma^2 B^2$ , and the characteristic polarization. The radiation itself had been predicted classically by G. A. Schott in 1912 and seen as a blinding blue-white glow at General Electric's 70-MeV synchrotron in 1947, where it was at first an unwanted energy drain on the accelerator. Schwinger's achievement was to show that this laboratory nuisance obeyed clean, universal formulae. The astrophysical leap came almost at once from others: Hannes Alfvén and Nicolai Herlofson proposed in 1950 that cosmic radio sources shine by this mechanism, Iosif Shklovsky argued in 1953 that the Crab Nebula's optical continuum and polarization were synchrotron light, and Vitaly Ginzburg and Sergei Syrovatskii built it into the standard theory of cosmic-ray electrons and the radio sky. What began as wasted beam energy in a physics lab became the lens through which we read magnetic fields and particle acceleration across the universe.

History

1912

— Schott publishes the classical theory of radiation from circular electron orbits

1947

— synchrotron light seen directly as a blue-white glow at the General Electric 70-MeV synchrotron

1949

— Schwinger derives the full classical angular and spectral distribution

1950

— Alfvén and Herlofson propose synchrotron emission for cosmic radio sources

1953

— Shklovsky identifies the Crab Nebula's optical light and polarization as synchrotron

1965

— Ginzburg and Syrovatskii codify cosmic-ray-electron synchrotron theory for the radio sky

Going further

Texts. Rybicki & Lightman, Radiative Processes in Astrophysics, Ch. 6 (the definitive treatment) · Longair, High Energy Astrophysics, 3rd ed., Ch. 8 · Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., §4.3 and Ch. 25 (beaming, nonthermal emission) · Ginzburg & Syrovatskii, The Origin of Cosmic Rays (the astrophysical foundation) · Dodelson, Modern Cosmology and Mo, van den Bosch & White, Galaxy Formation and Evolution (synchrotron foregrounds and the magnetized cosmos).

Provenance. Equations cross-checked against Rybicki & Lightman (Ch. 6), Longair (Ch. 8), and Carroll & Ostlie (§4.3). Standard constants: $\sigma_T=6.65\times10^{-25}\,\mathrm{cm^2}$ , $\nu_{B,0}\approx2.8\,\text{MHz}\,(B/\text{G})$ , $m_p/m_e\approx1836$ . A verifier should recompute: (1) the exponent algebra $\alpha=(p-1)/2$ with $p=2\Rightarrow\alpha=0.5$ , $p=3\Rightarrow\alpha=1.0$ ; (2) the scaling $\nu_c\propto\gamma^2 B$ and the $3/2$ prefactor, and the anchor $\nu_c\approx1.0$ GHz for $\gamma=5\times10^3$ , $B=10\,\mu$ G; (3) the $\nu^{5/2}$ optically-thick slope; (4) the $\Delta\alpha=1/2$ cooling break and $\nu_b\propto B^{-3}t^{-2}$ ; (5) the polarization ceiling $(p+1)/(p+7/3)=69\%$ for $p=2$ ; (6) the cooling time $t_{\rm cool}\approx50$ Myr for $\gamma=5\times10^3$ , $B=10\,\mu$ G.

Prerequisites. Specific Intensity and Radiative Flux · The Equation of Radiative Transfer · Optical Depth · Emission and the Source Function · Blackbody Radiation and the Planck Law. Leads to. Opacity and Absorption Coefficients · Kirchhoff's Laws and Thermal Emission · Stefan-Boltzmann and Wien Displacement Laws.

Ask the tutor: Ask me to walk through how a measured radio spectrum — its slope, its turnover, and its break — can be inverted into the electron energies, magnetic field, and age of a real source like Cassiopeia A or the M87 jet.

Why is synchrotron radiation called "nonthermal"?

Because it is not produced by particles in thermal (Maxwellian) equilibrium at a temperature. Thermal emission — like blackbody radiation — reflects a Maxwell-Boltzmann energy distribution and produces the smooth Planck curve. Synchrotron radiation instead comes from a power-law population of electrons, $N(\gamma)\propto\gamma^{-p}$ , that has been accelerated far out of equilibrium at shocks or in reconnection sites. That non-equilibrium energy distribution imprints a power-law spectrum $I_\nu\propto\nu^{-\alpha}$ with no characteristic temperature, and it is often polarized. When you see a power-law radio spectrum rising toward low frequency, or strong linear polarization, you are almost certainly looking at nonthermal synchrotron light rather than anything thermal.

How does measuring the spectral index tell us about particle acceleration?

The optically-thin slope is locked to the electron energy distribution by $\alpha=(p-1)/2$ , so measuring $\alpha$ gives you $p=2\alpha+1$ directly. Acceleration theories predict $p$ : diffusive shock acceleration at a strong non-relativistic shock predicts $p\approx2$ , hence $\alpha\approx0.5$ , and this is exactly what supernova remnants and many jets show. A steeper index means either intrinsically steeper injection or an aged, cooled population; a flatter one can indicate a hard spectrum or self-absorption. Because the mapping runs both ways, the radio spectrum becomes a remote probe of the microphysics of cosmic accelerators you could never sample in situ — you read the particle physics off the light.

What is synchrotron self-absorption and why does it give ν^(5/2)?

The same electrons that emit synchrotron photons can reabsorb them, and at low frequency, where the source is dense and the photons are soft, this self-absorption dominates. Once optically thick, the emergent intensity approaches the source function $S_\nu=j_\nu/\alpha_\nu$ , which behaves like Rayleigh-Jeans radiation at an effective temperature set by the energy of the electrons radiating at that frequency. Since those electrons have $\gamma\propto(\nu/B)^{1/2}$ , the effective temperature scales as $\nu^{1/2}$ , and the Rayleigh-Jeans form $S_\nu\propto\nu^2 T_{\rm eff}$ gives $\nu^2\cdot\nu^{1/2}=\nu^{5/2}$ . Crucially this slope is independent of $p$ , so a rising $\nu^{5/2}$ low-frequency spectrum is the unmistakable signature of a compact, self-absorbed synchrotron core.

Why does the spectrum steepen by exactly one half above the cooling break?

Because higher-energy electrons radiate away their energy faster — the cooling time scales as $1/\gamma$ — so above a break energy the population is depleted. Under continuous injection with radiative losses, the steady-state electron distribution above the break steepens by exactly one power of $\gamma$ , from $\gamma^{-p}$ to $\gamma^{-(p+1)}$ . Feeding that through $\alpha=(p-1)/2$ turns the photon index from $(p-1)/2$ into $p/2$ , a change of exactly $\Delta\alpha=1/2$ . The half-integer jump is model-independent for continuous injection, which is what makes it such a reliable diagnostic: when you see a spectrum break by $0.5$ , you are watching radiative aging, and the break frequency $\nu_b\propto B^{-3}t^{-2}$ dates the plasma.

If a phase-space process can give n < 1, can synchrotron electrons exceed c?

No — the electrons are always sub-luminal, with $\beta<1$ and finite $\gamma$ ; nothing in synchrotron radiation involves faster-than-light particles. What can look paradoxical is apparent superluminal motion in jets, where blobs of synchrotron-emitting plasma moving toward us at $\beta\to1$ appear on the sky to cross transverse distances faster than $c$ . That is a projection-and-light-travel-time illusion, not a real superluminal speed: the emitting material nearly catches up to its own light, so successive images arrive bunched in time and the apparent transverse speed can exceed $c$ while the true speed stays below it. It is the same $(1-\beta)$ pile-up that shortens the single-electron pulse and sets $\nu_c$ , seen on macroscopic scales.

How do we estimate the magnetic field from synchrotron observations?

The most common route is the equipartition (or minimum-energy) argument. The synchrotron luminosity depends on both the number of relativistic electrons and the field strength — more electrons or a stronger field both brighten the source — so a single flux measurement cannot separate them. Equipartition breaks the degeneracy by assuming the source arranges itself near the minimum total energy, which occurs when the particle and magnetic energy densities are comparable. That condition fixes $B$ from the observed luminosity, spectrum, and source size, typically giving microgauss fields in radio lobes and tens to hundreds of microgauss in supernova remnants. Independent checks come from the self-absorption turnover frequency and from X-ray inverse-Compton measurements, which together with the radio synchrotron flux pin down $B$ without assuming equipartition.