Synchrotron Self-Absorption

Point a radio telescope at a quasar core, a blazar jet, or a young compact radio galaxy and you meet the same puzzling shape: the spectrum climbs steeply from low frequencies, peaks in the GHz band, then falls away as a gentle power law. The falling side is ordinary optically-thin synchrotron radiation from a power-law sea of relativistic electrons — its slope reads off the electron energy index directly. The rising side is stranger: it rises as $\nu^{5/2}$ , a slope that is the same for every source no matter what the electrons are doing, and steeper than any thermal blackbody. That universal rise is the fingerprint of synchrotron self-absorption: at low frequencies the very electrons that emit the radiation reabsorb it, and the source becomes opaque to its own light. The physics that makes this tractable is a beautiful sleight of hand. A relativistic electron gas has no single temperature, so a blackbody argument seems hopeless. But the transfer problem only needs a source function $S_\nu=j_\nu/\alpha_\nu$ , and that ratio reads as a Rayleigh–Jeans brightness at an effective temperature set by whichever electrons radiate at $\nu$ . Because a synchrotron electron radiates near $\nu\propto\gamma^2 B$ , the one responsible for $\nu$ carries energy $\gamma m_e c^2\propto\nu^{1/2}$ , so the effective temperature scales as $\nu^{1/2}$ ; feed that into the $\nu^2$ of Rayleigh–Jeans and out drops $\nu^{5/2}$ . From the radiative-transfer equation this article builds the coefficients of a power-law plasma, the universal $5/2$ law, the turnover, and the way it weighs a distant field, caps the brightness temperature at $\sim10^{12}\ \mathrm{K}$ , and reveals relativistic beaming in jets.

Key takeaways

Synchrotron self-absorption (SSA) is the reabsorption of synchrotron photons by the same relativistic electrons that emit them; it dominates at low frequencies where the source goes optically thick to its own radiation.
Because relativistic electrons are non-thermal, the source function is set by an effective temperature $k_B T_{\rm eff}\simeq\gamma m_e c^2$ — the energy of the electrons radiating at that frequency — not a thermodynamic one.
The optically-thick spectrum rises as $S_\nu\propto\nu^{5/2}$ , independent of $p$ , turns over at $\nu_a$ , and joins the optically-thin power law $S_\nu\propto\nu^{-(p-1)/2}$ above it.
The turnover $\nu_a$ , peak flux $S_{\nu_a}$ , and angular size $\theta$ give the magnetic field $B\propto\nu_a^5\theta^4 S_{\nu_a}^{-2}(1+z)^{-1}$ — a rare equipartition-free way to weigh a distant field.
The intrinsic brightness temperature is capped near $10^{12}\ \mathrm{K}$ by the inverse-Compton catastrophe; apparent values above it signal Doppler beaming toward us.
Key idea: for a synchrotron source the source function is set not by a thermodynamic temperature but by the energy of the electrons radiating at each frequency; since that energy scales as $\nu^{1/2}$ , the optically-thick spectrum universally rises as $S_\nu\propto\nu^{5/2}$ , and the turnover frequency $\nu_a$ (where $\tau_\nu\sim1$ ) encodes the magnetic field and source size.

Where this lives today

Self-absorption is the daily working tool of very-long-baseline interferometry. A "flat" AGN spectrum is really the superposition of many self-absorbed components along the jet, each peaking at a lower $\nu_a$ as the field and density fall outward; the frequency-dependent position of the apparent core — the "core shift" — reconstructs the field profile $B(r)$ . The MOJAVE program has monitored hundreds of blazar jets at 15 GHz for two decades, combining self-absorption diagnostics with superluminal motions to map Doppler factors across the population. The brightness ceiling has itself been tested to breaking point: the space-VLBI mission RadioAstron measured apparent brightness temperatures above $10^{13}$ K in several quasars — so high that either the Doppler factors are extreme or the inverse-Compton limit needs refining, an open problem. At the opposite extreme the Event Horizon Telescope images the self-absorbed jet bases of M87 and Sgr A* at horizon scale, where the turnover sits in the millimetre band because those regions are the most compact and strongly magnetized. And in gigahertz-peaked sources the anti-correlation between turnover frequency and linear size marks them as the youngest radio galaxies.

The physical picture

Imagine standing inside a cloud of relativistic electrons threaded by a magnetic field. Each electron spirals and radiates a beamed cone of synchrotron photons. A photon of frequency $\nu$ heading outward can be reabsorbed by another electron — the inverse of emission — with a probability that is exactly what $\alpha_\nu$ measures. Lower the frequency and $\alpha_\nu$ climbs steeply, until the mean free path drops below the source size and a photon born deep inside cannot escape: the source is optically thick, and an outside observer sees only its "skin," the layer one optical depth deep. That skin does not shine like a blackbody at a single temperature — the gas is not thermal — but like a body at the effective temperature of the electrons dominating emission there. Because $\nu\propto\gamma^2 B$ , that electron has $\gamma\propto(\nu/B)^{1/2}$ , so $T_{\rm eff}\propto\nu^{1/2}$ , and the opaque brightness $2k_B T_{\rm eff}\,\nu^2/c^2$ becomes $\nu^2\times\nu^{1/2}=\nu^{5/2}$ . That single extra half-power over the thermal $\nu^2$ is the entire signature of self-absorption.

Figure

A synchrotron source at two frequencies. At low ν a photon escapes only from a thin τ = 1 skin, so we see a surface at the effective temperature of the emitting electrons; at high ν the source is transparent and we see the whole volume. Electrons radiating at ν = 1 GHz in B = 100 μG have γ ≈ 1900, energy ≈ 1 GeV, and T_eff ≈ 10¹³ K

Above the turnover the whole source is visible and the spectrum reverts to the electron power law; below it we see only the opaque skin and it climbs as $\nu^{5/2}$ . The branches meet at $\nu_a$ (where $\tau_{\nu_a}\sim1$ ), a turnover that is at once thermometer and ruler, tying together $\nu_a$ , angular size, and field. The second figure is the observable itself.

Figure

The self-absorbed synchrotron spectrum on log–log axes. Below the turnover the source is optically thick and rises with the universal slope S_ν ∝ ν^(5/2); above it the source is optically thin and falls as S_ν ∝ ν^(−0.75) for electron index p = 2.5. The branches meet at the self-absorption frequency ν_a ≈ 1 GHz where τ ≈ 1

The brightness temperature is not open-ended. When the core is compact and bright, its own synchrotron photons become a dense target for the same electrons, which up-scatter them by the inverse-Compton effect and drain their energy catastrophically. This "Compton catastrophe" caps the intrinsic brightness near $10^{12}\ \mathrm{K}$ ; sources that appear hotter are not defying physics — the emitting plasma streams toward us at relativistic speed, Doppler-boosting the apparent brightness.

The core equations

The whole treatment sits on the formal solution of the radiative-transfer equation for a uniform slab, with source function $S_\nu=j_\nu/\alpha_\nu$ and optical depth $\tau_\nu$ . Everything specific to synchrotron radiation enters through how $j_\nu$ and $\alpha_\nu$ depend on frequency and field.

I_\nu=S_\nu\left(1-e^{-\tau_\nu}\right) \qquad (1)

where: $I_\nu$ — emergent specific intensity; $S_\nu=j_\nu/\alpha_\nu$ — source function; $\tau_\nu=\alpha_\nu L$ — optical depth through path length $L$ ; $j_\nu,\ \alpha_\nu$ — emission and absorption coefficients. Reading it: for $\tau_\nu\ll1$ , $I_\nu\approx j_\nu L$ (thin); for $\tau_\nu\gg1$ , $I_\nu\approx S_\nu$ (thick). The SSA spectrum is the crossover. The emitting electrons are not thermal but follow a power law in energy, the one assumption that fixes both emission and absorption.

N(\gamma)\,\mathrm{d}\gamma=C\,\gamma^{-p}\,\mathrm{d}\gamma \qquad (2)

where: $N(\gamma)\,\mathrm{d}\gamma$ — electron density with Lorentz factor in $[\gamma,\gamma+\mathrm{d}\gamma]$ ; $C$ — normalization; $p$ — electron energy index, typically $2\lesssim p\lesssim3$ ; $\gamma$ — Lorentz factor. Reading it: $p$ is the one free number of the electron gas, and the miracle of SSA is that it vanishes from the optically-thick slope. An electron of Lorentz factor $\gamma$ radiates over a band peaking near a characteristic frequency set by $\gamma$ and the field — the hinge of the whole argument, converting electron energy into frequency.

\nu\simeq\gamma^2\,\nu_B\,\sin\alpha,\qquad \nu_B=\frac{eB}{2\pi m_e c} \qquad (3)

where: $\nu$ — frequency radiated by an electron of Lorentz factor $\gamma$ ; $\nu_B$ — electron gyrofrequency in cgs ( $\approx2.8\ \mathrm{MHz}$ per gauss); $B$ — field strength; $\alpha$ — pitch angle; $e,\ m_e,\ c$ — electron charge, mass, speed of light. Reading it: inverted, $\gamma\propto(\nu/B)^{1/2}$ — the electron responsible for a frequency has energy growing as $\nu^{1/2}$ , the seed of the $5/2$ slope and the effective-temperature picture. Integrating the single-electron power over the distribution gives the emissivity, whose slope defines the optically-thin index $\alpha=(p-1)/2$ .

j_\nu\propto C\,B^{(p+1)/2}\,\nu^{-(p-1)/2} \qquad (4)

where: $j_\nu$ — synchrotron emission coefficient; $(p-1)/2$ — optically-thin spectral index; $C,B$ — electron normalization and field. Reading it: a steeper electron distribution gives a steeper radio spectrum — how spectral indices become electron energy indices. The absorption coefficient comes from the same distribution via the stimulated-emission/true-absorption balance, and falls with frequency faster than the emissivity by $\nu^{-5/2}$ — why low frequencies go opaque first.

\alpha_\nu\propto C\,B^{(p+2)/2}\,\nu^{-(p+4)/2} \qquad (5)

where: $\alpha_\nu$ — self-absorption coefficient; $(p+4)/2$ — its steep frequency index; the extra half-power of $B$ relative to $j_\nu$ comes from the phase-space factor. Reading it: since $\alpha_\nu$ steepens faster than $j_\nu$ , the optical depth grows without bound toward low frequency — every synchrotron source is self-absorbed somewhere. Dividing emission by absorption gives the source function, and here the electron index $p$ and normalization $C$ both cancel, leaving the universal law.

S_\nu=\frac{j_\nu}{\alpha_\nu}\propto B^{-1/2}\,\nu^{5/2} \qquad (6)

where: $S_\nu$ — source function; $\nu^{5/2}$ — universal optically-thick slope, independent of $p$ ; $B^{-1/2}$ — a stronger field lowers the opaque brightness at fixed $\nu$ . Reading it: the $B^{-1/2}$ is counterintuitive but correct — a stronger field lets cooler electrons radiate there, so brightness drops. The same result follows without writing $j_\nu$ or $\alpha_\nu$ , by assigning each frequency the effective temperature of its emitting electrons — the physical heart of the phenomenon.

k_B T_{\rm eff}\simeq\gamma m_e c^2\propto\nu^{1/2}B^{-1/2} \qquad (7)

where: $T_{\rm eff}$ — effective temperature, the electron's kinetic energy over $k_B$ ; $k_B$ — Boltzmann constant. Reading it: substituting into $S_\nu=2k_B T_{\rm eff}\,\nu^2/c^2$ , the $\nu^{1/2}$ times $\nu^2$ gives $\nu^{5/2}$ — the same slope and $B^{-1/2}$ . The effective temperature at the turnover is essentially the observed brightness temperature, which cannot rise without limit — inverse-Compton losses impose a ceiling.

T_b=\frac{c^2 I_\nu}{2k_B\nu^2},\qquad T_b^{\max}\sim10^{12}\ \mathrm{K} \qquad (8)

where: $T_b$ — brightness temperature, defined so $I_\nu$ equals the Rayleigh–Jeans intensity of a body at $T_b$ ; the ceiling is the inverse-Compton limit. Reading it: an intrinsic $T_b$ far above $10^{12}\ \mathrm{K}$ radiates away its electron energy in a fraction of a light-crossing time, so persistent higher values must be Doppler-boosted. Finally, the turnover overdetermines the source: $\nu_a$ , the peak flux $S_{\nu_a}$ , and the size $\theta$ solve for the field with no equipartition assumption.

B\propto\nu_a^{5}\,\theta^{4}\,S_{\nu_a}^{-2}\,(1+z)^{-1} \qquad (9)

where: $\nu_a$ — turnover frequency; $\theta$ — angular diameter; $S_{\nu_a}$ — peak flux density; $z$ — redshift. Reading it: the large exponents demand accurate VLBI sizes and well-sampled turnovers, but in hand this is one of the very few model-independent field measurements in astrophysics.

Derivation AThe universal five-halves optically-thick spectrum

The goal of this derivation is to prove that the optically-thick branch of every synchrotron source rises as $\nu^{5/2}$ , regardless of the electron index $p$ or the normalization $C$ , by two independent routes — dividing the emission coefficient by the absorption coefficient, and assigning each frequency an effective electron temperature. Their agreement marks the result as a genuine property of the transfer problem, not an artifact of one calculational path.

Step 1. Write the emissivity, $j_\nu\propto C\,B^{(p+1)/2}\,\nu^{-(p-1)/2}$ , from integrating the single-electron power over $N(\gamma)=C\gamma^{-p}$ . Why: this fixes the numerator and gives the optically-thin slope $-(p-1)/2$ .

Step 2. Write the absorption coefficient, $\alpha_\nu\propto C\,B^{(p+2)/2}\,\nu^{-(p+4)/2}$ , from the stimulated-emission/true-absorption integral. Why: relative to $j_\nu$ it carries one extra $B^{1/2}$ and a steeper $\nu^{-5/2}$ — the origin of everything that follows.

Step 3. Form the ratio $S_\nu=j_\nu/\alpha_\nu$ : in frequency $\nu^{(p+4)/2-(p-1)/2}=\nu^{5/2}$ ; in field $B^{(p+1)/2}/B^{(p+2)/2}=B^{-1/2}$ ; and $C$ cancels. Why: both $p$ and the density $C$ disappear, so the opaque brightness is a universal function of frequency and field alone — the hallmark of SSA.

Step 4. Cross-check via effective temperature: $S_\nu=2k_B T_{\rm eff}\,\nu^2/c^2$ with $k_B T_{\rm eff}\simeq\gamma m_e c^2$ and $\gamma\propto\nu^{1/2}B^{-1/2}$ gives $S_\nu\propto\nu^{1/2}\cdot\nu^2\cdot B^{-1/2}=\nu^{5/2}B^{-1/2}$ . Why: the two routes agree exactly — the opaque source is a Rayleigh–Jeans surface at the emitting electrons' energy. *Verified: $(p+4)/2-(p-1)/2=5/2$ and field ratio $B^{-1/2}$ ; effective-temperature route gives identical slope and field dependence; matches Rybicki & Lightman §6.8 and Longair ch. 8.*

Go deeper · Derivation B · 1 minWhy the absorption coefficient steepens faster than the emissivity

Derivation BWhy the absorption coefficient steepens faster than the emissivity

This derivation traces where the extra $\nu^{-5/2}$ in the absorption coefficient comes from, so the $5/2$ law does not rest on a quoted formula. The self-absorption coefficient of a non-thermal gas is set not by the number of absorbers but by the balance of true absorption and stimulated emission, which for a power law reduces to an integral over the energy derivative of the occupation — exposing the two ingredients, a $\nu^{-2}$ phase-space factor and the power-law energy dependence, that produce the steep $(p+4)/2$ index.

Step 1. Start from the isotropic absorption coefficient $\alpha_\nu\propto\nu^{-2}\int E^2\,\dfrac{\partial}{\partial E}\!\left[\dfrac{N(E)}{E^2}\right]P(\nu,E)\,\mathrm{d}E$ , with $P(\nu,E)$ the single-electron power. Why: the $\nu^{-2}$ prefactor is a phase-space (photon-density) factor common to any stimulated process.

Step 2. Insert $N(E)\propto E^{-p}$ , so $N(E)/E^2\propto E^{-(p+2)}$ , its derivative $\propto E^{-(p+3)}$ , and the $E^2$ weight brings the integrand to $\propto E^{-(p+1)}P(\nu,E)$ . Why: the derivative of the occupation, not the occupation, sets the absorption — a quantum-statistical feature absent from the classical emissivity.

Step 3. With $\nu\propto\gamma^2 B$ the dominant contribution comes from $\gamma\propto(\nu/B)^{1/2}$ ; carrying the scalings through gives $\alpha_\nu\propto C\,B^{(p+2)/2}\,\nu^{-(p+4)/2}$ . Why: this is steeper than $j_\nu$ by exactly $\nu^{-5/2}$ , which makes $\tau_\nu$ rise toward low frequency and forces every source into self-absorption.

Step 4. Compare field powers: $j_\nu\propto B^{(p+1)/2}$ but $\alpha_\nu\propto B^{(p+2)/2}$ , one extra half-power. Why: the ratio for $S_\nu$ then leaves $B^{-1/2}$ — the same bookkeeping behind the counterintuitive fall of opaque brightness with field. *Verified: the $\nu^{-2}$ factor plus the derivative of the $E^{-(p+2)}$ occupation reproduce index $(p+4)/2$ ; consistent with Rybicki & Lightman eqs. 6.50 and 6.53.*

Go deeper · Derivation C · 1 minThe self-absorption frequency and the magnetic field

Derivation CThe self-absorption frequency and the magnetic field

The aim here is to turn the turnover into a measurement. Defining $\nu_a$ by optical depth of order unity, and using that the peak brightness equals the source function, lets us eliminate the two unobservable quantities — the normalization $C$ and the depth $L$ — in favor of three we can measure: the turnover frequency, the peak flux, and the angular size. The resulting field estimate assumes nothing about equipartition, which is why it stays a benchmark against energy-minimization arguments.

Step 1. Impose the turnover condition $\tau_{\nu_a}=\alpha_{\nu_a}L\simeq1$ , with $\tau_\nu\propto C\,B^{(p+2)/2}\,\nu^{-(p+4)/2}L$ . Why: this quantifies "the source turns opaque" and links the interior quantities $C,L$ to the observable $\nu_a$ .

Step 2. At the turnover the emergent intensity is close to the source function, $I_{\nu_a}\approx S_{\nu_a}\propto B^{-1/2}\nu_a^{5/2}$ . Why: the peak sits where the source crosses from thick to thin, so its brightness is the opaque-limit brightness.

Step 3. Convert flux and solid angle into brightness, $S_{\nu_a}=\dfrac{c^2}{2k_B}\dfrac{S_{\nu_a}^{\rm flux}}{\Omega\,\nu_a^2}$ with $\Omega\simeq\dfrac{\pi}{4}\theta^2$ . Why: an angular size $\theta$ turns the measured flux into the surface brightness that the source function represents.

Step 4. Equate the two expressions for $S_{\nu_a}$ , eliminate $C$ and $L$ , and solve: $B\propto\nu_a^{5}\,\theta^{4}\,\left(S_{\nu_a}^{\rm flux}\right)^{-2}\,(1+z)^{-1}$ . Why: the field now depends only on observables; redshift enters through the frequency shift and angular-size dimming, and the prefactor is an order-unity function of geometry and $p$ . *Verified: reproduces the standard equipartition-free field estimate (Slish 1963; Pacholczyk 1970; Marscher 1983 form) up to a geometry constant of order unity.*

Limiting cases

A homogeneous self-absorbed source has three regimes, stitched together at the turnover:

Case	Result	Meaning
Optically thin ( $\nu\gg\nu_a$ , $\tau_\nu\ll1$ )	$S_\nu\propto\nu^{-(p-1)/2}$	The intrinsic power law; the spectral index $\alpha=(p-1)/2$ reads off the electron energy index $p$ directly.
Optically thick ( $\nu\ll\nu_a$ , $\tau_\nu\gg1$ )	$S_\nu\propto\nu^{5/2}$	Universal rising slope, independent of $p$ — the fingerprint of SSA, and steeper than the thermal $\nu^2$ .
Turnover ( $\nu\approx\nu_a$ , $\tau_\nu\sim1$ )	$\tau_{\nu_a}\simeq1$	The spectral peak; $\nu_a$ , $S_{\nu_a}$ , and $\theta$ together yield $B$ and the electron density $C$ .
Strong field, fixed $\nu$	$S_\nu\propto B^{-1/2}$	A stronger field lets cooler electrons radiate at $\nu$ , so the opaque brightness falls.

Common trapTreating the ν^(5/2) rise as the thermal Rayleigh–Jeans ν² slope, or confusing SSA with free–free absorption

The optically-thick synchrotron slope is $\nu^{5/2}$ , not the blackbody Rayleigh–Jeans $\nu^{2}$ . The extra half-power is not a rounding error; it is the frequency dependence of the effective electron temperature, $T_{\rm eff}\propto\nu^{1/2}$ , and the single most diagnostic feature of the phenomenon — do not model a self-absorbed source as a blackbody at one temperature, because there is none. A second, more damaging confusion is with free–free (thermal bremsstrahlung) absorption by an external ionized screen: that has a different frequency dependence, $\alpha_\nu^{\rm ff}\propto\nu^{-2.1}$ , imprints a turnover with an underlying $\nu^{2}$ (not $\nu^{5/2}$ ) rise, and is a property of foreground gas, not of the emitting electrons — told apart by the rising-branch slope and how the turnover scales with source size.

Worked examples

1 · The electrons that make a GHz self-absorbed core (fully worked)

A jet component has $B=100\ \mu\mathrm{G}=10^{-4}\ \mathrm{G}$ and is observed at $\nu=1\ \mathrm{GHz}$ . Find the Lorentz factor, energy, and effective temperature of the electrons dominating emission there — and see why self-absorbed cores brush the brightness-temperature ceiling.

Step 1. The electron gyrofrequency is $\nu_B=eB/(2\pi m_e c)\approx2.8\ \mathrm{MHz}$ per gauss, so at $B=10^{-4}\ \mathrm{G}$ , $\nu_B=\mathbf{2.8\times10^{2}\ Hz}$ — the non-relativistic spiral rate, which the $\gamma^2$ boost carries up to the observed GHz.

Step 2. From $\nu\simeq\gamma^2\nu_B$ (taking $\sin\alpha\sim1$ ), $\gamma^2=\nu/\nu_B=10^{9}/2.8\times10^{2}\approx3.6\times10^{6}$ , so $\gamma\approx\mathbf{1.9\times10^{3}}$ . The electron energy is $E=\gamma m_e c^2=1900\times0.511\ \mathrm{MeV}\approx\mathbf{9.7\times10^{2}\ MeV}\approx1\ \mathrm{GeV}$ .

Step 3. The effective temperature is that energy over $k_B$ : $k_B T_{\rm eff}=E=9.7\times10^{2}\ \mathrm{MeV}=1.55\times10^{-3}\ \mathrm{erg}$ , so $T_{\rm eff}=1.55\times10^{-3}/1.38\times10^{-16}\approx\mathbf{1.1\times10^{13}\ K}$ .

Step 4. Interpret. A GHz self-absorbed core is opaque out to a skin at $T_{\rm eff}\sim10^{13}\ \mathrm{K}$ , near the $10^{12}\ \mathrm{K}$ inverse-Compton ceiling — and it is this near-coincidence that makes brightness temperature such a sharp beaming diagnostic. γ ≈ 1900, E ≈ 1 GeV, T_eff ≈ 10¹³ K — GHz self-absorbed cores sit near the inverse-Compton limit.

2 · Brightness temperature of a blazar core and its Doppler factor (fully worked)

VLBI resolves a blazar core into a component of flux $S_\nu=1\ \mathrm{Jy}$ at $\nu=5\ \mathrm{GHz}$ , size $\theta=0.2\ \mathrm{mas}$ , redshift $z=0.5$ . Find its rest-frame brightness temperature and a lower bound on the Doppler factor.

Step 1. For a uniform disk of angular diameter $\theta$ , $\Omega=\pi\theta^2/4$ and $T_b=\dfrac{c^2}{2k_B\nu^2}\dfrac{S_\nu}{\Omega}$ . Collecting unit conversions gives the working form $T_b\approx1.8\times10^{12}\,\left(\dfrac{S_\nu}{\mathrm{Jy}}\right)\left(\dfrac{\nu}{\mathrm{GHz}}\right)^{-2}\left(\dfrac{\theta}{\mathrm{mas}}\right)^{-2}(1+z)\ \mathrm{K}$ .

Step 2. Substitute: $\left(\nu/\mathrm{GHz}\right)^{2}=25$ , $\left(\theta/\mathrm{mas}\right)^{2}=0.04$ , so the geometric factor is $S_\nu/(\nu^2\theta^2)=1/(25\times0.04)=1.0$ . Then $T_b\approx1.8\times10^{12}\times1.0\times(1+0.5)=\mathbf{2.6\times10^{12}\ K}$ .

Step 3. Compare with the ceiling. The intrinsic inverse-Compton limit is $T_b^{\max}\sim10^{12}\ \mathrm{K}$ , and the observed value exceeds it by a factor $\approx2.6$ . Since a moving source boosts brightness temperature by one power of the Doppler factor, $T_b^{\rm obs}=\delta\,T_b^{\rm int}$ , the field must be beamed with $\delta\gtrsim T_b^{\rm obs}/T_b^{\max}\approx\mathbf{2.6}$ .

Step 4. Sanity check. A Doppler factor of a few is exactly what superluminal motion and rapid variability independently imply for blazar jets — the brightness-temperature excess and the apparent superluminal speeds are two faces of the same beaming. T_b ≈ 2.6×10¹² K exceeds the 10¹² K ceiling, requiring δ ≳ 2.6.

Physics of the portrait

Synchrotron self-absorption is the meeting point of accelerator physics and the radio sky: the classical theory of radiation from relativistically gyrating electrons, worked out for particle accelerators, turned skyward once it was clear the strongest cosmic radio sources shine by the same mechanism — and reabsorb their own light at low frequencies.

Origins · 1 minVitaly Ginzburg and how this was discovered

Vitaly Ginzburg

1916–2009 · Soviet theory of cosmic synchrotron radiation and self-absorption

Vitaly Lazarevich Ginzburg was one of the architects of the synchrotron interpretation of cosmic radio emission. In the early 1950s, following Alfvén and Herlofson's suggestion and Shklovsky's identification of the Crab Nebula's continuum as synchrotron light, Ginzburg and his collaborator Sergei Syrovatskii built the systematic theory of how a power-law sea of relativistic electrons in a magnetic field radiates across the radio spectrum — including the regime where the source becomes opaque to itself. Their treatment of self-absorption, alongside contemporaneous work by Le Roux and Slish, showed that the low-frequency turnover of compact sources is intrinsic to the emitting plasma rather than an external screen, so the turnover encodes the field strength and source size. Ginzburg's breadth was extraordinary — from the Ginzburg–Landau theory of superconductivity (for which he shared the 2003 Nobel Prize) to the origin of cosmic rays — and the self-absorption formalism he helped establish is still the tool VLBI observers reach for when they weigh a distant jet's field.

History

1947

— first visible synchrotron light is seen at the General Electric 70 MeV electron synchrotron

1950

— Alfvén and Herlofson, and independently Kiepenheuer, propose synchrotron radiation for cosmic radio sources

1953

— Shklovsky identifies the Crab Nebula continuum as synchrotron emission

1963

— Slish, and Ginzburg and Syrovatskii, formalize synchrotron self-absorption and its low-frequency turnover

1969

— Kellermann and Pauliny-Toth establish the ~10¹² K inverse-Compton brightness-temperature limit

Going further

Texts. Rybicki & Lightman, Radiative Processes in Astrophysics, ch. 6 (synchrotron) and §6.8 (self-absorption) · Longair, High Energy Astrophysics, 3rd ed., ch. 8 · Carroll & Ostlie, An Introduction to Modern Astrophysics, ch. 9 and 28 (radiative transfer; active galaxies) · Pacholczyk, Radio Astrophysics (1970), self-absorption formalism · Ryden, Introduction to Cosmology, radiation-processes background.

Provenance. The scalings $j_\nu\propto\nu^{-(p-1)/2}$ , $\alpha_\nu\propto\nu^{-(p+4)/2}$ , and $S_\nu\propto B^{-1/2}\nu^{5/2}$ are verified against Rybicki & Lightman §6.3–6.8 (eqs. 6.36, 6.50, 6.53) and Longair ch. 8; the characteristic frequency $\nu\simeq\gamma^2\nu_B\sin\alpha$ with $\nu_B=eB/2\pi m_e c\approx2.8\ \mathrm{MHz}$ per gauss is standard cgs. Worked example 1: $\nu_B=280\ \mathrm{Hz}$ at $100\ \mu\mathrm{G}$ gives $\gamma\approx1.9\times10^3$ , $E\approx1\ \mathrm{GeV}$ , $T_{\rm eff}\approx1.1\times10^{13}\ \mathrm{K}$ . Worked example 2: $T_b\approx1.8\times10^{12}\,(S/\mathrm{Jy})(\nu/\mathrm{GHz})^{-2}(\theta/\mathrm{mas})^{-2}(1+z)\ \mathrm{K}$ gives $T_b\approx2.6\times10^{12}\ \mathrm{K}$ , $\delta\gtrsim2.6$ . The $10^{12}\ \mathrm{K}$ limit is Kellermann & Pauliny-Toth (1969). A verifier should recompute $(p+4)/2-(p-1)/2=5/2$ , the effective-temperature route $T_{\rm eff}\propto\nu^{1/2}B^{-1/2}$ , and the field exponents $B\propto\nu_a^5\theta^4 S^{-2}(1+z)^{-1}$ (Slish 1963; Pacholczyk 1970; Marscher 1983).

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

Ask the tutor: Want me to walk through inverting a measured turnover — $\nu_a$ , peak flux, and VLBI size — into an actual magnetic field and Doppler factor for a real blazar core?

Why is the optically-thick slope 5/2 and not the thermal 2?

Because the emitting gas is not thermal — there is no single temperature. The transfer problem only needs the source function $S_\nu=j_\nu/\alpha_\nu$ , and for a synchrotron plasma that ratio behaves like a Rayleigh–Jeans brightness at an effective temperature: the energy of the electrons dominating emission at $\nu$ , over $k_B$ . Those electrons have $\gamma\propto\nu^{1/2}$ (since $\nu\propto\gamma^2 B$ ), so $T_{\rm eff}\propto\nu^{1/2}$ , and multiplying the Rayleigh–Jeans $\nu^2$ by this extra $\nu^{1/2}$ gives $\nu^{5/2}$ — the frequency dependence of the effective temperature made visible.

Does the self-absorbed slope really not depend on the electron index p?

No — that is what makes it useful. The emissivity carries $\nu^{-(p-1)/2}$ and the absorption coefficient $\nu^{-(p+4)/2}$ ; dividing them, the exponent difference is $(p+4)/2-(p-1)/2=5/2$ for every $p$ , and $C$ cancels too, so the rising branch is a universal $\nu^{5/2}$ whether the electrons are steep or flat. The optically-thin branch is the opposite — its slope $-(p-1)/2$ depends entirely on $p$ , so observers read the electron energy index off the falling side, not the rising side.

How can the optically-thick brightness go down when the field goes up?

The source function is $S_\nu\propto B^{-1/2}\nu^{5/2}$ , so at fixed frequency a stronger field gives a fainter opaque surface. The reason is $\nu\propto\gamma^2 B$ : at fixed $\nu$ a larger $B$ means the radiating electrons have a smaller Lorentz factor, $\gamma\propto(\nu/B)^{1/2}$ , hence lower energy, lower effective temperature, and a dimmer Rayleigh–Jeans surface. A stronger field recruits cooler electrons there, so only the opaque-limit brightness falls, not the total emitted power.

What actually causes the turnover — is the source running out of photons?

No; it is running out of transparency. As frequency drops, $\alpha_\nu\propto\nu^{-(p+4)/2}$ climbs steeply, so $\tau_\nu=\alpha_\nu L$ reaches and then exceeds unity. Below that a photon born deep inside is reabsorbed before it escapes, and we see only the thin skin one optical depth deep — fewer photons reach us not because fewer are made but because most are recaptured. The turnover $\nu_a$ is simply where $\tau_{\nu_a}\simeq1$ , the boundary between seeing the whole source and seeing only its surface.

Why does the brightness temperature stop at about 10¹² K?

When a self-absorbed core is compact and bright, its own synchrotron photons fill it densely enough that the same electrons inverse-Compton scatter them to higher energies. The scattered power climbs extremely fast with brightness temperature — roughly as $T_b^5$ once a second scattering generation kicks in — so above about $10^{12}\ \mathrm{K}$ the electrons radiate away their energy in far less than a light-crossing time and cool back below the limit almost instantly. This runaway is the "Compton catastrophe"; apparent values above the ceiling are real photons Doppler-boosted by bulk relativistic motion toward us.

How is SSA distinguished from free–free absorption, which also makes a turnover?

Both bend a spectrum over at low frequency, but differ in slope, origin, and scaling. SSA is intrinsic to the emitting electrons and gives a rising branch of $\nu^{5/2}$ ; free–free absorption is caused by an external screen of ionized gas and gives an underlying $\nu^{2}$ rise with $\alpha_\nu^{\rm ff}\propto\nu^{-2.1}$ . Because SSA depends on the source's own size and field, its turnover scales tightly with angular size (the $\nu_a$ – $\theta$ relation of GPS sources), whereas a free–free turnover depends on foreground plasma. The rising slope and the size correlation usually separate the two cleanly.