Active Galactic Nuclei: Unified Model

For half a century the centres of active galaxies were a taxonomist's nightmare. Seyfert galaxies split into Type 1 and Type 2 by whether their optical spectra carried broad emission lines; quasars, radio galaxies, BL Lac objects, and flat-spectrum radio quasars each earned a separate name from a separate discovery. By the 1980s the literature carried more than a dozen classes, each apparently a distinct beast. The unified model is the claim that most of this zoo is one animal seen from different sides. The physical core is always the same: a supermassive black hole of $10^6$ – $10^{10}\,M_\odot$ accreting through a disk that radiates the bulk of the luminosity, a fast broad-line region (BLR) within a fraction of a parsec, a slower narrow-line region (NLR) at tens to hundreds of parsecs, and — in some objects — a relativistic jet. Wrapped around the disk at parsec scales sits a geometrically and optically thick dusty torus. What changes is not the engine but our sight line. Look down the axis and you see the naked continuum and broad lines (Type 1); look through the torus and dust extinguishes the BLR, leaving only narrow lines and warm infrared (Type 2). Two further knobs — jet power (radio-loud versus radio-quiet) and accretion rate relative to Eddington — complete the scheme, so a radio-loud jet swinging toward us runs radio galaxy → broad-line radio galaxy → quasar → blazar.

Key takeaways

The diverse AGN classes — Seyfert 1 and 2, radio-loud and radio-quiet quasars, broad- and narrow-line radio galaxies, blazars — are largely one object: a black hole accreting through a disk inside a dusty obscuring torus.
The primary axis is orientation: a sight line inside the torus opening reveals the broad-line region and X-ray continuum (Type 1); a sight line through the torus hides them, leaving only narrow lines (Type 2).
Two more axes refine it: jet power (radio-loud versus radio-quiet) and accretion rate $\dot m=\dot M/\dot M_{\rm Edd}$ (thin disk versus advection-dominated flow).
The Eddington luminosity $L_{\rm Edd}\approx1.26\times10^{38}(M_\bullet/M_\odot)\ \mathrm{erg\,s^{-1}}$ sets the power scale; the Seyfert ratio $N_2/N_1\approx4$ fixes $\theta_{\rm tor}\approx37^\circ$ and covering fraction $C_f=\cos\theta_{\rm tor}\approx0.80$ .
Blazars are the extreme-orientation case: a jet at $\theta\lesssim1/\gamma$ is boosted by $\delta^{3+\alpha}\sim10^5$ for $\gamma=10$ , so each blazar implies $\sim2\gamma^2\approx200$ misaligned parent radio galaxies.
Orientation is a framework, not a total theory: changing-look AGN, true Type 2s, and the receding torus require genuine intrinsic evolution.
Key idea: a single accretion-powered engine plus an anisotropic obscuring torus, viewed from different angles, reproduces most AGN spectral classes; orientation, jet power, and accretion rate are the three governing parameters.

Where this lives today

Orientation unification is now a working tool, and its frontier is where pure geometry breaks down. Infrared interferometry with the VLTI (MIDI and GRAVITY) has resolved the parsec-scale dust in nearby Seyferts, showing the "torus" is less a smooth doughnut than a clumpy structure with a polar dusty wind — a refinement, not a refutation. Hard X-ray surveys from Swift/BAT and NuSTAR measure line-of-sight columns directly and find that much of the accretion in the Universe is Compton-thick ( $N_H>1.5\times10^{24}\ \mathrm{cm^{-2}}$ ), hidden even in deep optical surveys — the absorption-corrected census that closes the cosmic X-ray background. Reverberation campaigns like SDSS-RM weigh thousands of black holes across cosmic time through the $R_{\mathrm{BLR}}$ – $L$ and virial relations above. Meanwhile changing-look AGN and the IceCube neutrino association of the flaring blazar TXS 0506+056 in 2017 keep the field honest: the unified model organizes the zoo, but accretion physics, jet launching, and feedback are where the open questions now live.

The physical picture

Picture a brilliant lamp — the accretion disk — at the centre of a thick ring of smoke, the dusty torus. Stand above the ring and you see the naked lamp and the fast motes of dust swirling beside it: the broad-line region, gas orbiting at thousands of kilometres per second within a fraction of a parsec, its lines Doppler-broadened into smooth humps. Stand in the plane of the ring and the smoke blocks both. Now you see only distant motes, lit by light escaping through the central hole — the narrow-line region at tens of parsecs — plus the ring itself glowing warm in the infrared. Same lamp, same motes, two utterly different spectra. The first figure fixes this geometry and reads a real number off it.

Figure

One engine, two sight lines. A black hole, its accretion disk, the fast broad-line clouds, and the parsec-scale dusty torus, wrapped by narrow-line gas and a jet. A sight line down the axis θ < θ_tor sees the broad-line region as a Type 1 nucleus; a sight line through the torus hides it, leaving a Type 2. For an observed Seyfert ratio N₂/N₁ ≈ 4 the opening half-angle is θ_tor ≈ 37° and the covering fraction is C_f = cosθ_tor ≈ 0.80

The ionization cone is the searchlight beam. Beyond hiding the nucleus, the torus collimates the disk's ionizing radiation into two opposing cones, so the narrow-line gas lights up only where the beam reaches it. Directly imaging these biconical [O III] structures in nearby Seyferts — half-opening angles of $30$ – $50^\circ$ — is a geometric fingerprint of the torus, independent of any single spectrum. Relativistic beaming is the amplifier. A jet moving at bulk Lorentz factor $\gamma\sim10$ toward us has its emission boosted by the Doppler factor to the third or fourth power — a factor of order $10^5$ — so an intrinsically ordinary radio galaxy, if its jet points within roughly $1/\gamma$ of the line of sight, becomes a hyper-variable, superluminal blazar. The second figure shows how steeply that boost falls off as the jet swings away, and why blazars must be rare.

Figure

Relativistic Doppler boost of a jet blob for bulk Lorentz factor γ = 10 and spectral index α = 1: the observed flux scales as δ⁴ with δ = 1 divided by γ(1 − β cosθ). Pointing straight at us θ = 0 gives δ ≈ 20 and a boost of 1.6×10⁵; at the beaming-cone edge θ = 1/γ ≈ 5.7° the boost is still 10⁴; beyond θ ≈ 25° the jet is de-boosted below unity

The two-axis picture — orientation (Type 1 versus Type 2) crossed with jet power (radio-loud versus radio-quiet) — organizes almost the entire AGN zoo into a $2\times2$ grid, with luminosity and Eddington ratio as continuous refinements within each cell.

The core equations

The scheme rests on a small set of relations: a power scale, a size scale, a pair that convert line widths into masses, and a pair governing beaming and obscuration. The first is the Eddington luminosity, where radiation pressure on the accreting plasma cancels gravity.

L_{\mathrm{Edd}} = \frac{4\pi G M_{\bullet} m_p c}{\sigma_T} \approx 1.26\times10^{38}\left(\frac{M_{\bullet}}{M_\odot}\right)\ \mathrm{erg\,s^{-1}} \qquad (1)

where: $M_{\bullet}$ — black hole mass; $m_p$ — proton mass; $\sigma_T$ — Thomson cross section; $c$ — speed of light. Reading it: the maximum steady luminosity depends only on mass, so a $10^8\,M_\odot$ black hole cannot long outshine $\sim1.3\times10^{46}\ \mathrm{erg\,s^{-1}}$ — a bright quasar. The engine's size is set by the gravitational radius. The inner edge of a radiatively efficient disk sits at a few $r_g$ , and the accretion continuum is emitted within tens of $r_g$ .

r_g = \frac{G M_{\bullet}}{c^2}, \qquad r_g \approx 1.5\times10^{9}\left(\frac{M_{\bullet}}{10^6 M_\odot}\right)\ \mathrm{m} \qquad (2)

where: $r_g$ — gravitational radius; $M_{\bullet}$ — black hole mass. Reading it: for $10^6\,M_\odot$ the scale is $\sim1.5\times10^9\ \mathrm{m}$ (a few solar radii); for $10^8\,M_\odot$ it grows to $\sim1.5\times10^{11}\ \mathrm{m}\approx1\ \mathrm{AU}$ — smaller than the Solar System, yet outshining a whole galaxy. The luminosity is drawn from gravitational binding energy released as gas spirals inward; the radiative efficiency $\eta$ converting rest-mass flux into light dwarfs any nuclear-burning yield.

L_{\mathrm{bol}} = \eta\,\dot{M}c^2, \qquad \lambda_{\mathrm{Edd}} \equiv \frac{L_{\mathrm{bol}}}{L_{\mathrm{Edd}}} \qquad (3)

where: $L_{\mathrm{bol}}$ — bolometric luminosity; $\eta$ — radiative efficiency, $\approx0.057$ (Schwarzschild) to $\approx0.42$ (maximal Kerr); $\dot{M}$ — mass accretion rate; $\lambda_{\mathrm{Edd}}$ — Eddington ratio. Reading it: at $\eta\approx0.1$ a black hole converts ten percent of $\dot M c^2$ to light — twenty times the efficiency of hydrogen fusion — and $\lambda_{\mathrm{Edd}}$ says how hard it drives relative to its limit. The broad-line clouds sit at a radius set by photoionization balance, so their distance scales as the square root of luminosity — the relation exploited by reverberation mapping.

R_{\mathrm{BLR}} \propto L^{1/2}, \qquad v_{\mathrm{BLR}} \sim \sqrt{\frac{G M_{\bullet}}{R_{\mathrm{BLR}}}} \sim \text{few}\times10^3\ \mathrm{km\,s^{-1}} \qquad (4)

where: $R_{\mathrm{BLR}}$ — BLR radius, from the light-travel lag between continuum and line variations; $L$ — ionizing continuum luminosity; $v_{\mathrm{BLR}}$ — virial velocity setting the broad-line width. Reading it: brighter nuclei push the BLR outward as $L^{1/2}$ , so the lag pins the radius without resolving it. Combining the lag radius with the line width gives the most productive black hole weighing method beyond the local Universe — the virial (reverberation) mass.

M_{\bullet} = f\,\frac{R_{\mathrm{BLR}}\,\Delta v^2}{G} \qquad (5)

where: $f$ — order-unity geometry factor for the unknown BLR inclination and structure; $\Delta v$ — broad-line velocity width; $R_{\mathrm{BLR}}$ — reverberation-mapped radius. Reading it: the mass is a virial product $R\,\Delta v^2/G$ , with $f$ calibrated against the local $M_\bullet$ – $\sigma$ relation; it drives essentially every quasar black hole census. Blazar physics turns on the relativistic Doppler factor — how strongly a jet's emission is enhanced by its motion toward us.

\delta = \frac{1}{\gamma(1-\beta\cos\theta)} \qquad (6)

where: $\delta$ — Doppler factor; $\gamma$ — bulk Lorentz factor; $\beta$ — jet speed in units of $c$ ; $\theta$ — angle to the line of sight. Reading it: head-on and ultrarelativistic, $\delta\to2\gamma$ , so a $\gamma=10$ jet reaches $\delta\approx20$ . That factor enters the observed flux three times, plus once through the spectral slope, giving the beaming law that makes blazars extreme.

S_{\nu}^{\mathrm{obs}} = \delta^{3+\alpha}\, S_{\nu}^{\mathrm{emit}} \qquad (7)

where: $S_\nu$ — flux density; $\alpha$ — spectral index, $S_\nu\propto\nu^{-\alpha}$ ; $\delta$ — Doppler factor. The exponent is $3+\alpha$ for a discrete blob, $2+\alpha$ for a continuous jet. Reading it: for $\gamma=10$ , $\theta=0$ , $\alpha=1$ the boost is $20^4=1.6\times10^5$ — intrinsic luminosity is irrelevant next to geometry. Finally, if engines point randomly, the ratio of obscured to unobscured objects fixes the torus opening angle and covering fraction.

\frac{N_{\mathrm{Type\,2}}}{N_{\mathrm{Type\,1}}} = \frac{\cos\theta_{\mathrm{tor}}}{1-\cos\theta_{\mathrm{tor}}} = \frac{C_f}{1-C_f}, \qquad C_f = \cos\theta_{\mathrm{tor}} \qquad (8)

where: $\theta_{\mathrm{tor}}$ — torus opening half-angle from the polar axis; $C_f=\cos\theta_{\mathrm{tor}}$ — covering fraction, the sky fraction blocked as seen from the engine; $N$ — number of each type. Reading it: a narrower opening (larger $\cos\theta_{\mathrm{tor}}$ ) blocks more sight lines and raises the Type 2 excess.

Derivation AThe Eddington luminosity as a force balance

The goal is to pin down the single luminosity at which an accreting black hole holds itself in check — the point where the outward push of radiation on the ionized plasma exactly cancels the inward pull of gravity, so that steady spherical accretion can no longer proceed. We balance the two forces on a coupled proton–electron pair, exploit the fact that both fall as $r^{-2}$ so the balance is scale-free, and read off a luminosity that depends on black hole mass alone.

Step 1. The outward force on an electron is the radiation momentum flux times its cross section: $F_{\mathrm{rad}}=\sigma_T L/(4\pi r^2 c)$ . Why: photons carry momentum $E/c$ , and electrons dominate the scattering because $\sigma_T\propto m^{-2}$ , so the light pushes on the electrons, not the protons.

Step 2. The inward force is gravity on the proton, which drags the electron with it: $F_{\mathrm{grav}}=G M_\bullet m_p/r^2$ . Why: nearly all the plasma mass is in the protons, so the effective gravitating mass per scattering electron is one proton mass — the pair moves as one neutral fluid element.

Step 3. Set them equal; the common $r^{-2}$ cancels, leaving $L_{\mathrm{Edd}}=4\pi G M_\bullet m_p c/\sigma_T$ . Why: radius-independence is the crucial fact — above this luminosity radiation pressure beats gravity at every radius at once, so the flow is unbound everywhere and steady accretion halts.

Step 4. Insert constants: $4\pi G m_p c/\sigma_T=6.3\times10^4\ \mathrm{erg\,s^{-1}\,g^{-1}}$ , and $\times M_\odot=1.99\times10^{33}\ \mathrm{g}$ gives $L_{\mathrm{Edd}}\approx1.26\times10^{38}(M_\bullet/M_\odot)\ \mathrm{erg\,s^{-1}}$ . Why: the natural power unit of AGN — a $10^8\,M_\odot$ black hole near Eddington reaches $\sim10^{46}\ \mathrm{erg\,s^{-1}}$ , the luminosity of luminous quasars. *Why the result is clean: inserting cgs constants gives $4\pi G m_p c/\sigma_T=1.26\times10^{38}\ \mathrm{erg\,s^{-1}}\,M_\odot^{-1}$ exactly, and the $r^{-2}$ dependence cancels precisely between radiation pressure and gravity, confirming that the balance holds at every radius simultaneously.*

Go deeper · Derivation B · 1 minType 2 to Type 1 counts and the torus covering fraction

Derivation BType 2 to Type 1 counts and the torus covering fraction

The goal is to turn a headcount into a geometry: if active nuclei are identical engines pointing in random directions, the ratio of obscured (Type 2) to unobscured (Type 1) objects fixes the torus opening angle and covering fraction, with no free parameter beyond that single angle. We assume isotropic orientation, integrate the sight-line probability over the polar opening cones, and invert the observed Seyfert ratio to read the geometry directly — the quantitative backbone of orientation unification.

Step 1. For isotropic axes, the probability that the observer's inclination lies in $[i,i+di]$ is $P(i)\,di=\tfrac12\sin i\,di$ on $[0,\pi]$ , normalized to unity. Why: a random axis distributes the sight line uniformly over the sphere, weighted by solid angle $\sin i$ — more sky near the equator than the pole.

Step 2. We see the broad-line region (Type 1) only if the sight line falls inside one of the two polar opening cones, $i<\theta_{\mathrm{tor}}$ or $i>\pi-\theta_{\mathrm{tor}}$ , giving $P_1=1-\cos\theta_{\mathrm{tor}}$ . Why: each cone contributes $\tfrac12(1-\cos\theta_{\mathrm{tor}})$ , and there are two — one per pole — since the torus is symmetric top and bottom.

Step 3. Everything else is obscured, so $P_2=1-P_1=\cos\theta_{\mathrm{tor}}$ , which is precisely the covering fraction: $C_f=\cos\theta_{\mathrm{tor}}$ . Why: covering fraction is the probability a random sight line is blocked; a narrower opening (smaller $\theta_{\mathrm{tor}}$ , larger $\cos\theta_{\mathrm{tor}}$ ) blocks more, so $C_f\to1$ as the opening closes.

Step 4. The observable ratio is $N_2/N_1=\cos\theta_{\mathrm{tor}}/(1-\cos\theta_{\mathrm{tor}})$ . Inverting $N_2/N_1\approx4$ gives $\cos\theta_{\mathrm{tor}}=0.8$ , hence $\theta_{\mathrm{tor}}\approx37^\circ$ , $C_f\approx0.80$ . Why: a belt covering $\sim80\%$ of sight lines matches the imaged $30$ – $50^\circ$ ionization-cone opening angles — two independent measures of one geometry. *Why the numbers are self-consistent: at $\theta_{\mathrm{tor}}=45^\circ$ , $\cos45^\circ=0.707$ gives $N_2/N_1=0.707/0.293\approx2.4$ — a narrower opening is needed to push the ratio toward the observed $3$ – $4$ . Equivalently, $C_f=\cos\theta_{\mathrm{tor}}=N_2/(N_1+N_2)$ , so a ratio of $4$ gives $C_f=0.8$ , self-consistent with the $80\%$ covering fraction.*

Go deeper · Derivation C · 1 minRelativistic beaming and the blazar fraction

Derivation CRelativistic beaming and the blazar fraction

The goal is to explain why a jet that is intrinsically unremarkable becomes, when it happens to point at us, the most extreme class of active galaxy — and to quantify both the boost and the rarity at once. We track how the Doppler factor $\delta$ enters the observed flux through three distinct relativistic transformations plus the spectral slope, then use the forward cone of half-angle $1/\gamma$ to compute what fraction of jets we catch beamed, tying blazars to their misaligned parent radio galaxies.

Step 1. For a jet at angle $\theta$ , $\delta=1/[\gamma(1-\beta\cos\theta)]$ ; at $\theta=0$ with $\beta\to1$ this reaches $\delta\to2\gamma$ , since $1-\beta\approx1/2\gamma^2$ . Why: aberration and arrival-time compression combine into one kinematic factor that head-on is twice the Lorentz factor — a $\gamma=10$ jet gives $\delta\approx20$ .

Step 2. The flux picks up three powers of $\delta$ : aberration compresses emission forward, $d\Omega^{\mathrm{obs}}=\delta^{-2}d\Omega^{\mathrm{emit}}$ ; photon energy blueshifts by $\delta$ ; arrival rate speeds by $\delta$ . Since $I_\nu/\nu^3$ is Lorentz-invariant, $I_\nu^{\mathrm{obs}}=\delta^3 I_\nu^{\mathrm{emit}}$ ; folding in $S_\nu\propto\nu^{-\alpha}$ gives $S_\nu^{\mathrm{obs}}=\delta^{3+\alpha}S_\nu^{\mathrm{emit}}$ for a discrete blob. Why: each factor of $\delta$ is a distinct effect, and the exponent $3+\alpha$ (versus $2+\alpha$ for a steady jet) fingerprints the geometry.

Step 3. Numerically, $\gamma=10$ , $\theta=0$ , $\alpha=1$ gives $\delta=20$ and a boost $20^4=1.6\times10^5$ . Why: a factor $10^5$ dwarfs any spread in intrinsic jet power, so a modest radio galaxy pointed at us masquerades as a superluminal blazar.

Step 4. Strong boosting needs $\theta\lesssim1/\gamma$ ; the chance a random jet lies in that cone is $f_{\mathrm{blazar}}=1-\cos(1/\gamma)\approx1/(2\gamma^2)=0.5\%$ for $\gamma=10$ . Why: each beamed blazar implies $\sim2\gamma^2\approx200$ misaligned parents — why blazars are rare, and the origin of the pairing BL Lacs ↔ FR I, FSRQs ↔ FR II. *Why the numbers close: $\delta(0)\approx2\gamma=20$ and $20^4=1.6\times10^5$ ; the beaming-cone solid-angle fraction evaluates to $1-\cos(0.1\,\mathrm{rad})=4.99\times10^{-3}\approx1/(2\gamma^2)$ , and $2\gamma^2=200$ , confirming both the boost and the parent-population count.*

Limiting cases

The model is really a map from a few parameters — inclination $i$ , jet angle $\theta_{\mathrm{jet}}$ , and Eddington-scaled accretion rate $\dot m$ — onto the observed classes. The limiting regimes make that map explicit.

Case	Result	Meaning
Pole-on, radio-quiet	$i \lesssim \theta_{\mathrm{tor}}$	Sight line inside the torus opening: direct view of the continuum, X-ray corona, and broad-line region gives a Seyfert 1 or Type 1 quasar.
Edge-on, radio-quiet	$i \gtrsim \theta_{\mathrm{tor}}$	Sight line through the torus: BLR and soft X-rays absorbed, only narrow lines survive — a Seyfert 2 / Type 2, with hidden broad lines recoverable in polarized light.
Jet toward observer	$\theta_{\mathrm{jet}} \lesssim 1/\gamma$	Radio-loud AGN with the jet inside the beaming cone: Doppler-boosted, rapidly variable non-thermal continuum swamps the host — a blazar (BL Lac or FSRQ).
Low accretion rate	$\dot m = \dot M/\dot M_{\mathrm{Edd}} \lesssim 10^{-2}$	The thin disk is replaced by a hot advection-dominated flow; the BLR and torus weaken or vanish, giving LINER-like low-luminosity AGN that pure orientation cannot unify.

Common trapOrientation is the whole story

It is not. Orientation unifies a great deal, but three facts show intrinsic differences are also real. First, some "true Type 2" AGN show no hidden broad lines even in deep spectropolarimetry — at low accretion rate the BLR may not form, so there is no obscured Type 1 to reveal. Second, changing-look AGN switch type on timescales of months to years, far too fast for the torus to move across our sight line; the broad lines genuinely appear or vanish as the ionizing continuum changes. Third, the covering fraction falls with luminosity — the "receding torus" — as rising radiation pressure drives out the inner dust. Orientation explains geometry, not accretion-mode changes or intrinsic evolution; treating every Type 2 as a hidden Type 1 will mislead you.

Worked examples

1 · Power budget of a luminous quasar (fully worked)

A luminous quasar hosts a black hole of $M_\bullet=10^8\,M_\odot$ and shines at bolometric luminosity $L_{\mathrm{bol}}=10^{46}\ \mathrm{erg\,s^{-1}}$ . Find its Eddington luminosity, its Eddington ratio, and the mass accretion rate feeding it, taking a radiative efficiency $\eta=0.1$ .

Step 1. The Eddington luminosity follows from equation (1): $L_{\mathrm{Edd}}=1.26\times10^{38}\times10^8=\mathbf{1.26\times10^{46}\ erg\,s^{-1}}$ . This is the ceiling a $10^8\,M_\odot$ black hole can sustain against radiation pressure.

Step 2. The Eddington ratio is $\lambda_{\mathrm{Edd}}=L_{\mathrm{bol}}/L_{\mathrm{Edd}}=10^{46}/1.26\times10^{46}=\mathbf{0.79}$ — the quasar radiates at nearly $80\%$ of its limit, a near-critical accretor typical of the brightest quasars.

Step 3. Invert the accretion-luminosity relation (3): $\dot M=L_{\mathrm{bol}}/(\eta c^2)=10^{46}/[0.1\times(3.0\times10^{10})^2]=1.1\times10^{26}\ \mathrm{g\,s^{-1}}$ . Converting with $1\ \mathrm{yr}=3.16\times10^{7}\ \mathrm{s}$ and $M_\odot=1.99\times10^{33}\ \mathrm{g}$ gives $\dot M=\mathbf{1.8\ M_\odot\,yr^{-1}}$ — swallowing two suns a year, a tenth of it emerging as light, outshines the whole host galaxy. $L_{\mathrm{Edd}}=1.3\times10^{46}\ \mathrm{erg\,s^{-1}}$ , $\lambda_{\mathrm{Edd}}\approx0.8$ , $\dot M\approx1.8\ M_\odot\,\mathrm{yr^{-1}}$ .

2 · Weighing a black hole by reverberation (fully worked)

Monitoring a nearby Seyfert 1 shows its broad H $\beta$ line responding to continuum variations with a light-travel lag of $20$ days, and the line has a velocity width $\Delta v=3000\ \mathrm{km\,s^{-1}}$ . Estimate the black hole mass, taking the geometry factor $f\approx1$ .

Step 1. The lag is the light-crossing time to the BLR, so $R_{\mathrm{BLR}}=c\times20\ \mathrm{days}=3.0\times10^{10}\ \mathrm{cm\,s^{-1}}\times(20\times86400\ \mathrm{s})=\mathbf{5.2\times10^{16}\ cm}$ — about $0.017$ parsec, comfortably inside the parsec-scale torus, which is exactly why the BLR is what the torus hides.

Step 2. Apply the virial mass (5) with $\Delta v=3.0\times10^{8}\ \mathrm{cm\,s^{-1}}$ : $M_\bullet=f R_{\mathrm{BLR}}\Delta v^2/G=1\times(5.2\times10^{16})(3.0\times10^{8})^2/(6.67\times10^{-8})$ . The numerator is $5.2\times10^{16}\times9.0\times10^{16}=4.7\times10^{33}$ , so $M_\bullet=4.7\times10^{33}/6.67\times10^{-8}=7.0\times10^{40}\ \mathrm{g}$ .

Step 3. Convert to solar masses: $M_\bullet=7.0\times10^{40}/1.99\times10^{33}=\mathbf{3.5\times10^{7}\ M_\odot}$ . Cross-check: such a black hole has $L_{\mathrm{Edd}}=4.4\times10^{45}\ \mathrm{erg\,s^{-1}}$ , so at $\lambda_{\mathrm{Edd}}\sim0.1$ it shines near $10^{44.5}\ \mathrm{erg\,s^{-1}}$ — consistent with the modest luminosity a $20$ -day lag implies through the $R$ – $L$ relation. $M_\bullet\approx3.5\times10^{7}\ M_\odot$ .

3 · How rare are blazars? (fully worked)

A radio galaxy carries a jet with bulk Lorentz factor $\gamma=10$ and spectral index $\alpha=0.7$ . Compare the observed core flux when the jet points at us ( $\theta=0$ ) against when it lies at $\theta=30^\circ$ , and estimate how many such radio galaxies exist per beamed blazar.

Step 1. Pointed at us, $\delta=1/[\gamma(1-\beta)]\approx2\gamma=20$ . The boost is $\delta^{3+\alpha}=20^{3.7}$ . Since $\log_{10}20=1.301$ , $20^{3.7}=10^{4.81}=\mathbf{6.5\times10^{4}}$ .

Step 2. At $\theta=30^\circ$ , $\delta=1/[10(1-0.995\cos30^\circ)]=0.72$ , so the boost is $0.72^{3.7}=\mathbf{0.30}$ — the jet is de-boosted below its rest-frame flux. The ratio of orientations is $6.5\times10^4/0.30\approx\mathbf{2\times10^{5}}$ : the same jet is hundreds of thousands of times brighter aimed at us than $30^\circ$ away.

Step 3. Strong boosting needs $\theta\lesssim1/\gamma=5.7^\circ$ , a fraction $f_{\mathrm{blazar}}\approx1/(2\gamma^2)=\mathbf{0.5\%}$ . So per beamed blazar there are $\sim2\gamma^2\approx\mathbf{200}$ misaligned radio galaxies — why blazars are a small minority, and why their de-projected space density must match the numerous radio galaxies for unification to hold.

Physics of the portrait

The unified model is not a single equation but an argument won by one decisive observation, and the person most associated with turning a plausible geometric idea into a testable, tested claim is Robert Antonucci — whose spectropolarimetry of NGC 1068 with Joseph Miller in 1985 caught a Type 2 nucleus hiding a Type 1 in its polarized light.

Origins · 1 minRobert Antonucci and how this was discovered

Robert Antonucci

American astrophysicist, hidden broad-line discovery in NGC 1068, 1985 · the observational case for orientation-based AGN unification

By the early 1980s the idea that Seyfert 2 galaxies might be Seyfert 1s seen through obscuring material was in the air, but it was speculation without a decisive test. Robert Antonucci, with Joseph Miller at Lick Observatory, supplied the test in 1985. They took a polarized-light spectrum of the archetypal Seyfert 2 NGC 1068 and found, in the scattered (and therefore polarized) component, the broad permitted lines entirely absent from the direct total-light spectrum. The interpretation was immediate: a hidden Type 1 nucleus, its broad-line region blocked from direct view, was leaking broad-line photons past the torus, where free electrons and dust scattered a few into our sight line and polarized them. Seeing the hidden engine in a mirror was the most direct confirmation that Type 2 AGN harbour obscured Type 1 nuclei. Antonucci went on to write the definitive 1993 Annual Review, which framed the unified model for a generation and set the agenda — orientation first, intrinsic differences second — that the field still argues within today.

History

1908

— Fath records strong nuclear emission lines in the spiral NGC 1068

1943

— Seyfert catalogues galaxies with bright, broad-line nuclei

1963

— Schmidt identifies 3C 273 as a redshifted quasar, revealing quasars' vast luminosities

1985

— Antonucci and Miller find polarized broad lines hidden inside the Seyfert 2 NGC 1068

1995

— Urry and Padovani codify the orientation unification of radio-loud AGN

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 28 (galaxies and AGN) · Mo, van den Bosch & White, Galaxy Formation and Evolution, Ch. 14 (active galactic nuclei) · Krolik, Active Galactic Nuclei (Princeton, 1999) · Peterson, An Introduction to Active Galactic Nuclei · Binney & Tremaine, Galactic Dynamics, for the stellar-dynamical black hole mass foundations · Dodelson, Modern Cosmology, for the accretion and structure-growth context. Key numerical benchmarks worth keeping in mind: the Eddington coefficient $1.26\times10^{38}\ \mathrm{erg\,s^{-1}}\,M_\odot^{-1}$ follows from $4\pi G m_p c/\sigma_T$ ; the gravitational-radius scale is $r_g\approx1.5\times10^{9}\ \mathrm{m}$ for $10^6\,M_\odot$ (and $\approx1\ \mathrm{AU}$ for $10^8\,M_\odot$ ); the beaming exponent $3+\alpha$ arises from the three Doppler factors together with the invariance of $I_\nu/\nu^3$ ; the covering-fraction relation $C_f=\cos\theta_{\mathrm{tor}}=N_2/(N_1+N_2)$ comes from integrating $\int\tfrac12\sin i\,di$ ; and the blazar fraction is $\approx1/(2\gamma^2)$ . The quoted AGN values — ionization-cone opening angles $30$ – $50^\circ$ , Compton-thick threshold $1.5\times10^{24}\ \mathrm{cm^{-2}}$ , and radiative efficiencies $0.057$ – $0.42$ — are standard figures from the review literature.

Prerequisites. Supermassive Black Holes · Eddington Luminosity and Accretion Limits · Thin Accretion Disks and the Shakura–Sunyaev Model. Leads to. Blazars and Relativistic Beaming · Relativistic Jets and Launching Mechanisms · Quasars and Quasar Evolution · AGN Feedback.

Ask the tutor: Ask me to walk through why a single spectropolarimetric spectrum of NGC 1068 was enough to turn orientation unification from a guess into the standard picture.

Is the AGN unified model a complete theory of active galaxies?

No — it is a unifying framework, not a first-principles theory. It says much of the observed diversity is a projection effect: the same black hole, torus, and jet seen from different angles. That is very well supported for the orientation axis (Type 1 versus Type 2) and the beaming axis (radio galaxy versus blazar). But orientation cannot explain why some nuclei launch powerful jets and others do not, why the accretion rate spans orders of magnitude, why changing-look AGN switch type in months, or why the torus recedes with luminosity. The model organizes the phenomenology and hands the deeper questions — accretion mode, jet launching, feedback — to the physics of the engine itself.

What is the dusty torus actually made of, and how big is it?

It is a geometrically and optically thick distribution of gas and dust encircling the disk on parsec scales, from about a tenth of a parsec (set by dust sublimation at the hot inner edge) out to a few parsecs. Its defining property is anisotropy: it absorbs the optical and ultraviolet continuum along equatorial sight lines and re-emits it as warm infrared, while leaving polar sight lines clear. Modern infrared interferometry shows it is not a smooth doughnut but a clumpy structure, often with a dusty polar wind driven off the inner edge by radiation pressure. Read "torus" as shorthand for anisotropic circumnuclear obscuration, not a literal solid ring.

Why does a jet pointing at us make such an enormous difference?

Because relativistic beaming compounds several effects at once. A blob moving toward us at Lorentz factor $\gamma$ has its emission swept forward by aberration, blueshifted in frequency, and sped up in arrival rate — each a factor of the Doppler factor $\delta$ — so the flux scales as $\delta^{3+\alpha}$ . For $\gamma=10$ pointed at us, $\delta\approx20$ and the boost is about $10^5$ , swamping any spread in intrinsic jet power. A jet aimed within $\sim1/\gamma$ of the line of sight dominates the whole source and produces the hyper-variable, superluminal blazar phenomenon; swing it a few degrees away and the boost collapses.

How can we weigh a black hole in a galaxy we cannot resolve?

Through reverberation mapping and the virial method. The broad-line clouds orbit the black hole, so their velocity width $\Delta v$ reflects the depth of its potential. We cannot resolve the orbit, but we can time it: when the continuum brightens, the broad lines respond a light-travel time later, giving $R_{\mathrm{BLR}}=c\,\tau$ . Combining the two into the virial product $M_\bullet=f R_{\mathrm{BLR}}\Delta v^2/G$ , with a geometry factor $f$ calibrated against the local $M_\bullet$ – $\sigma$ relation, yields the mass. A single lag-plus-width measurement weighs a black hole millions of light-years away to within a factor of a few.

What is the difference between radio-loud and radio-quiet AGN?

The presence and power of a relativistic jet. Radio-loud AGN — roughly ten to twenty percent of the population — launch powerful collimated jets that dominate their radio emission and, when beamed at us, produce blazars. Radio-quiet AGN have weak or absent jets and radiate mostly from the disk and corona. The split correlates with host type (radio-loud sources favour massive ellipticals) and is thought to depend on black hole spin and the magnetic field threading the inner disk, which set the jet-launching efficiency. Radio loudness is thus a second, largely independent axis, crossed with orientation to give the full two-by-two organization.

Are quasars and Seyfert galaxies the same kind of object?

Essentially yes — they differ mainly in luminosity, and the boundary is a historical convention, not a physical divide. Seyferts are lower-luminosity AGN whose host galaxy is clearly visible; quasars are higher-luminosity AGN so bright the point-like nucleus outshines and hides the host, which is why they were first mistaken for stars ("quasi-stellar"). Both are accreting supermassive black holes with the same anatomy — disk, broad- and narrow-line regions, torus — and both come in Type 1 and Type 2 set by orientation. Turn up the accretion rate and mass and a Seyfert simply becomes a quasar; the physics is continuous across the name.

What is a changing-look AGN, and why does it matter?

One that transitions between Type 1 and Type 2 — its broad lines appearing or vanishing — over months to a few years. It matters because that is far too fast for orientation: the parsec-scale torus cannot drift across our sight line so quickly. Instead the broad-line region itself is switching on or off, most plausibly because the ionizing continuum from the inner disk is rising or crashing as the accretion rate changes. Changing-look sources are direct evidence that not every class difference is geometric — some are intrinsic and time-variable — and they are now a prime laboratory for accretion-state transitions in real time.