Active galaxies and quasars: the light of feeding supermassive black holes

At a glance

An active galactic nucleus (AGN) is the compact centre of a galaxy that shines because gas is falling into a supermassive black hole. The hot accretion disc can outshine all the stars of its galaxy; the most luminous AGN are called quasars. Seyfert galaxies, radio galaxies, quasars and blazars are now understood largely as one engine seen from different directions and at different power.

Sizes and line widths: typical ranges from Netzer (2015); Eddington luminosity computed from the formula in Fabian (2012). Records as of October 2026. pc = parsec (3.26 light-years); L☉, M☉ = solar luminosity and mass.

Sources: [1] [2] [3] [4] [5] [6] [7]

What makes a galaxy active?

Large galaxies generally host central black holes of millions to billions of solar masses (see our black holes page), but most, including the one in the Milky Way, are usually quiet. A nucleus becomes active when gas flows in and, spiralling towards the black hole, heats up and radiates a sizeable fraction of its rest-mass energy; an efficiency of about 10 per cent is a common working assumption. Netzer’s review counts a nucleus as an AGN when it shines above roughly 10⁻⁵ of its Eddington limit (explained below), and calls those above 1 per cent efficient, “radiative-mode” accretors.

The output spans the spectrum: the disc shines mostly in ultraviolet and visible light, surrounding dust re-emits in the infrared, X-rays come from hot gas close to the black hole, and jets produce radio waves and, when aimed at us, highly variable emission. AGN therefore turn up in optical, X-ray, infrared and radio surveys alike.

Sources: [1] [2] [8]

From Seyfert’s nuclei to the first quasar

In 1943 Carl Seyfert published spectra of six spiral galaxies with unusually bright nuclei, among them NGC 1068, NGC 4151 and NGC 7469. Their nuclei showed strong emission lines from highly ionised gas, broadened by Doppler motions of up to about 8,500 km/s; in NGC 4151 the hydrogen lines had narrow cores with wings spanning 7,500 km/s. Such objects are now called Seyfert galaxies.

Radio astronomy then turned up bright sources that looked like faint stars in optical images. In 1963 Maarten Schmidt recognised the puzzling emission lines in the spectrum of one of them, 3C 273, as hydrogen lines shifted to longer wavelengths by 15.8 per cent (redshift z = 0.158). The most direct conclusion, he wrote, was that the “star” was the enormously bright nucleus of a distant galaxy. These quasi-stellar sources, or quasars, proved to be the most luminous members of the AGN family.

Sources: [9] [3]

One engine, many faces: the unified model

The unified model explains most AGN with a few components around the black hole (Figure 1). Closest in is the accretion disc. Gas clouds moving at thousands of kilometres per second, 0.01 to 1 parsec out depending on luminosity, form the broad-line region. Beyond them lies a dusty torus, an axisymmetric structure 0.1 to 10 parsecs in size whose inner edge is set by the temperature at which dust grains evaporate, about 1,800 K. Radiation escaping along the axis ionises slower gas out to hundreds or thousands of parsecs, the narrow-line region, often seen as two cones. Some AGN also launch a pair of jets.

What we see then depends on the viewing angle. Looking past the torus, we see the bright central source with both broad and narrow lines: a type 1 object such as a Seyfert 1 or a quasar. When the torus blocks the view, only narrow lines remain: a type 2 object such as a Seyfert 2. Among radio-loud AGN, jets seen side-on belong to radio galaxies, while a jet pointing almost at us makes a rapidly variable, strongly polarised blazar. Urry and Padovani’s classic 1995 review links luminous (Fanaroff–Riley type II) radio galaxies with radio quasars, and less luminous type I radio galaxies with the blazars called BL Lac objects.

Key evidence came from polarised light. In 1985 Antonucci and Miller found broad lines hidden in the polarised spectrum of the Seyfert 2 galaxy NGC 1068: light from a concealed broad-line region was being scattered towards us. Deeper surveys found many type 2 objects without hidden lines, so not every type 2 may be a hidden type 1. In 2018 the GRAVITY instrument, combining the four large telescopes of ESO’s Very Large Telescope, resolved the broad-line region of 3C 273 at about 10 microarcseconds: the gas rotates around a black hole of about 300 million solar masses, with its rotation axis aligned with the jet.

Sources: [1] [8] [10]

Broad lines, narrow lines and weighing black holes

An AGN’s spectrum records this structure (Figure 2). Permitted lines such as hydrogen’s Hβ are broadened in the fast-moving broad-line region to 1,000–20,000 km/s, while the narrow-line region produces lines only 300–1,000 km/s wide. The narrow region also emits forbidden lines, such as the strong [O III] lines of doubly ionised oxygen, which form only in low-density gas; their absence from the broad component shows that the broad-line clouds are much denser.

The broad lines also weigh the black hole. Their brightness follows changes in the central continuum after a delay equal to the light-travel time across the broad-line region, typically about a light-month; this is reverberation mapping. Combining that radius with the line width, a measure of orbital speed, gives the enclosed mass, much as Kepler’s laws do for planets. GRAVITY’s resolved view of 3C 273 supports the method’s assumptions and offers an independent benchmark. The spectral laboratory shows how line shifts encode motion.

Sources: [1] [10]

The Eddington limit and how black holes grow

Radiation pushes on infalling gas. When the outward push of light on electrons balances gravity’s pull on protons, an AGN reaches the Eddington luminosity, L_Edd = 4πGMm_pc/σ_T: about 1.3 × 10³¹ watts, or some 33,000 Suns, for every solar mass of black hole. Quasars appear to respect this limit. With a radiative efficiency of 10 per cent, a black hole shining at its limit grows by a factor e (about 2.7) roughly every 50 million years.

That clock is a problem for the earliest quasars. J0313−1806, reported in 2021 at redshift 7.642, shines with about 3.6 × 10¹³ solar luminosities only 670 million years after the Big Bang, powered by a black hole of about 1.6 billion solar masses; by our estimate that is roughly 70 per cent of its Eddington luminosity. Growing such a mass from a 100-solar-mass seed at the Eddington rate would take about 17 e-foldings, some 800 million years, more than the age of the Universe at the time. Black holes must therefore have started from much heavier seeds, grown faster than the Eddington rate for a while, or both, a question discussed on our black holes page.

Sources: [2] [11]

Calculator: the Eddington limit

The calculator opens on a luminous quasar: a billion-solar-mass black hole shining at half its Eddington limit. Note how much matter it must swallow each year, then change the mass and the Eddington ratio, or compare the Seyfert and Sagittarius A* presets.

For the default quasar it gives an Eddington luminosity of about 1.3 × 10⁴⁰ W (3.3 × 10¹³ L☉), an accretion rate of about 11 solar masses per year and a growth time of 100 million years. The efficiency η is an input because it depends on the black hole’s spin. The calculator assumes ionised hydrogen; for gas of solar composition Netzer quotes a slightly higher limit, 1.5 × 10³¹ W per solar mass, and, as the notes say, accretion above the limit is also observed.

Sources: [1]

Jets, radio galaxies and blazars

Jets are narrow beams of plasma moving close to the speed of light, launched from near the black hole. In low-luminosity Fanaroff–Riley type I radio galaxies the radio emission peaks near the nucleus, while powerful type II sources have more highly collimated jets. Because the plasma moves so fast, its emission is beamed forward: a jet aimed at us looks far brighter and more variable than one seen side-on, and features in it can appear to move faster than light (superluminal motion).

The best-studied jet base is in the nearby galaxy M87, whose black hole the Event Horizon Telescope imaged at 1.3 mm (see our black holes page). Observations made in 2018 at 3.5 mm and published in 2023 captured the ring and the jet together: at that wavelength the ring is about 8.4 Schwarzschild radii across, roughly 50 per cent larger than at 1.3 mm, pointing to emission from the accretion flow itself, and the jet’s launching region is wider than expected for a jet powered by the black hole alone, suggesting an additional wind. See also radio astronomy and cosmic rays and neutrinos.

Sources: [8] [12]

Quasars across cosmic time

Quasars were far more common in the past than they are today: their activity peaked at redshifts of about 2–3, when galaxies were richest in gas. The Sloan Digital Sky Survey’s sixteenth quasar catalogue, released in 2020, lists 750,414 quasars, 225,082 of them appearing in an SDSS release for the first time. X-ray surveys, which are less affected by obscuring dust, add many hidden AGN that optical surveys miss.

Reaching the earliest epochs is harder, because the light of the most distant quasars is stretched into the infrared. J0313−1806 held the distance record from 2021 until 2026, when ESA’s Euclid found more distant quasars (see Current research). Even then, quasars contained black holes of a billion solar masses or more, key evidence on how the first black holes formed (see Big Bang and cosmic history).

Sources: [1] [2] [6] [11]

Feedback: how black holes shape galaxies

A black hole holds only a tiny fraction of its galaxy’s mass, yet the energy released as it grows, assuming 10 per cent efficiency, exceeds the gravitational binding energy of the galaxy’s bulge by a large factor. If even a small part couples to the surrounding gas, the AGN can heat or expel it, halting star formation and its own growth. This feedback is a leading explanation for why black-hole masses track the masses of their hosts and why massive galaxies stopped forming stars (see galaxies and the cosmic web).

Two modes are distinguished. In the radiative or quasar mode, winds driven by a luminous AGN push gas outwards; such outflows are hard to observe directly, though J0313−1806 shows absorbing gas moving at up to about 20 per cent of the speed of light. In the kinetic or radio mode, common in massive elliptical galaxies, jets inflate bubbles in the hot gas of galaxy clusters; in the Perseus cluster, the brightest in the X-ray sky, Chandra images show concentric ripples around such bubbles. This heating appears to balance the cooling of cluster gas over long periods.

Sources: [2] [11]

Current research (2024–2026)

The most luminous known quasar. In February 2024 a team led by Christian Wolf reported that J0529−4351 is the most luminous object known: it shines with more than 500 trillion times the Sun’s luminosity, from an accretion disc about seven light-years across, around a black hole of about 17 billion solar masses that grows by roughly one solar mass per day. Its light has travelled for over 12 billion years.

An X-ray census. The first data release of the eROSITA all-sky X-ray survey, published in January 2024 for half of the sky, contains more than 900,000 sources, about 710,000 of them supermassive black holes in distant galaxies.

A new distance record. In July 2026 the Euclid collaboration reported 31 new quasars at redshifts 6.6–7.8, selected from the first 1.5 years of the Euclid Wide Survey and confirmed with the Keck, Magellan and Large Binocular telescopes. Twelve lie at z ≥ 7; the most distant, EUCL J172902.75+641018.1 at z ≈ 7.77, beats the 2021 record (see our Euclid page).

Little red dots. The James Webb Space Telescope has revealed many compact, red early-Universe sources with broad hydrogen lines. A 2024 study of such objects at z ≈ 4–5.5 read their Hα lines, 1,200–3,700 km/s wide, as signs of black holes of 10⁷–10⁸ solar masses, about ten times more common than extrapolations of the quasar population. Their nature is debated. A study now published in Nature argues that they are young black holes wrapped in dense cocoons of ionised gas, where scattering by electrons broadens the lines; on that reading the masses are only 10⁵–10⁷ solar masses, about two orders of magnitude lower than earlier estimates.

Sources: [4] [7] [5] [13] [14]

Going further: data archives

AGN research leans heavily on public archives. The Sloan Digital Sky Survey releases optical spectra, including those of the SDSS-V Black Hole Mapper (Data Release 20 at the time of writing); NED, the NASA/IPAC Extragalactic Database, gathers measurements and literature for extragalactic objects; the Chandra Data Archive holds X-ray observations; the eROSITA-DE DR1 site provides the eRASS1 catalogues; the MOJAVE programme publishes Very Long Baseline Array monitoring of jets; and MAST serves Hubble and JWST data.

Known vs open questions

Established results

• AGN are powered by gas falling onto supermassive black holes.
• Orientation explains much of their variety, from type 2 nuclei with hidden broad lines to blazars.
• Broad-line gas orbits within about a parsec of the black hole, and GRAVITY has resolved it in 3C 273.
• Quasars appear to respect the Eddington limit and were most active at redshifts of about 2–3.
• Black holes of a billion solar masses existed less than 700 million years after the Big Bang.

Open questions

• How did the first supermassive black holes form and grow so quickly?
• Are all type 2 AGN hidden type 1s, and what sets how much of the sky the torus covers?
• How are jets launched, and how much AGN energy actually reaches the host galaxy?
• What are JWST’s little red dots, and how massive are their black holes?

Related on this site

• Research dossier: galaxy evolution
• Black holes
• Galaxies and the cosmic web
• Relativity and gravity
• Big Bang and cosmic history
• James Webb Space Telescope
• Spectral laboratory