Quasar Winds at 8,400 km/s Shut Down Early Galaxies, JWST Survey Confirms

Six of 27 quasars observed by the James Webb Space Telescope (JWST) at the cosmic dawn are firing galaxy-scale winds at speeds up to 8,400 kilometres per second, a detection rate roughly 6.6 times higher than in any equivalent sample from later in cosmic history. The new survey, published in Nature on 6 May, identifies what the lead authors call cosmic blowtorches: radiation-driven outflows powerful enough to scour star-forming gas out of an entire galaxy in a single quasar duty cycle.

The bigger payoff sits one layer below the headline. For three years, JWST has kept finding massive, red, quiescent galaxies less than a billion years after the Big Bang, structures that, on standard models, simply had not lived long enough to grow up and stop forming stars. The 22% extreme-outflow rate finally supplies the off switch.

Six Cosmic Blowtorches Caught in the Act

The survey targets the brightest quasars known at redshift 5 to 6, a sample of 27 objects drawn from roughly 100 spectroscopically confirmed sources with absolute magnitudes brighter than M₁₄₅₀ = -25.5. Each one was imaged with NIRSpec, JWST’s near-infrared integral field unit, which slices a galaxy into a grid of spectra and lets astronomers trace gas motion at every point in the frame.

The tracer of choice is the forbidden [O III] λ5007 emission line, a feature ionised oxygen produces in low-density gas far from the central black hole. Because the line cannot survive in the dense inner accretion zone, any blueshift it shows comes from gas at kiloparsec scales, the body of the host galaxy. Six of the 27 spectra show that gas streaming towards Earth at v98 values between 2,700 and 8,400 km/s.

“Many of those galaxies looked old in the sense that they had stopped forming stars long before it would be expected,” said Weizhe Liu, a postdoctoral scholar at the University of Arizona’s Steward Observatory and the paper’s first author. The team’s outflow census, he and co-author Xiaohui Fan argue in the University of Arizona research announcement, is the first that meets the energy threshold simulations have long demanded.

That threshold matters because most quasar outflows reported until now have been too feeble to do real damage. Five of the six new detections carry kinetic energy outflow rates between 10% and 260% of the host quasar’s bolometric luminosity, easily clearing the 0.1% to 5% range cosmological simulations say is needed to shut star formation down.

How Radiation Pressure Empties a Galaxy

A quasar is what an actively feeding supermassive black hole looks like from a great distance: an accretion disk so luminous it outshines the entire galaxy around it. The new work focuses on a feedback channel that is not the famous polar jet but the wider, slower wind the bright disk itself produces.

Radiation from the accretion disk pushes on dust and ionised gas through ordinary photon momentum. In early galaxies, which are denser and clumpier than the elegant spirals of today, that radiation pressure couples to gas across many directions at once rather than being collimated into a narrow beam. The result is a near-isotropic outflow that can spread well past the galaxy’s stellar body.

Mass, Speed, and Reach

The numbers in the Nature paper are unforgiving. Mass outflow rates run from 140 to 46,000 solar masses per year. One object’s wind carries roughly 1,000 solar masses of gas a year, similar to its inferred star-formation rate, meaning the quasar is removing fuel as fast as the galaxy can burn it.

At 8,400 km/s, the fastest wind in the sample easily exceeds galactic escape velocity. The gas does not just slosh out and fall back; it can be flung past the galaxy’s circumgalactic envelope into the intergalactic medium, where it cannot cool and condense back into stars on any cosmologically interesting timescale.

Why the Off Switch Sticks

Quenching only works if it is durable. Three properties of these outflows make them durable:

• Wide-angle geometry. Radiation-driven winds spray across most of the sky as seen from the black hole, so very little of the host galaxy is shielded.
• Energy budget. When kinetic energy reaches double-digit percentages of the quasar’s total luminosity, the wind heats and shocks gas faster than it can radiate the energy away.
• Duty cycle length. The team estimates extreme outflow quasars stay active for around 100 million years, long enough to depopulate a galaxy’s reservoir of cold gas.

After the central engine fades, the galaxy is left without raw material to make new stars. It cools, reddens, and looks ancient even when the universe around it is still young.

The Dead-Galaxy Puzzle JWST Created for Itself

From its first deep fields in 2022, JWST has kept producing observations that did not fit. The most stubborn category was a population of massive, passively evolving galaxies sitting at redshifts where they should still be busy assembling. A few hundred million years of cosmic time is not, in standard galaxy formation models, enough to build a Milky Way-mass stellar population and then turn off the spigot.

The orbiting observatory had detected the corpses without producing the murder weapon. Earlier candidate explanations leaned on feedback from supernovae, which struggle to expel enough gas from a galaxy as massive as the ones being observed, or on quasar jets, which are narrow and miss most of the host. Neither story matched the geometry or the energy required.

The new outflow survey closes that gap. A radiation-driven wind from a luminous early quasar carries enough mass, covers a wide enough solid angle, and runs for long enough to leave behind exactly the kind of stalled, gas-poor galaxy JWST has been finding. The mechanism is also consistent with another JWST puzzle: black holes in this epoch routinely look overweight relative to their host galaxies’ stellar mass, which is what happens when feedback chokes off the host’s growth while the central engine keeps gorging.

Why Quasar Winds Hit Harder at Dawn

The most surprising piece of the result is not the velocities themselves; broad absorption-line quasars at lower redshifts can match them. It is how often these extreme winds turn up. The team built two control samples matched in luminosity to the early sample, and the contrast is sharp.

The combined lower-redshift detection rate sits below 3%, against 22% in the early sample. The average kinetic-energy outflow rate at z 5 to 6 is roughly 100 times what is seen in the comparison populations.

Why the early universe in particular? Two factors compound. Gas reservoirs in early galaxies are far denser, so radiation couples to matter more efficiently and the wind picks up more momentum per square parsec. And the supermassive black holes themselves were already startlingly massive at this epoch, which means their accretion disks could pump out enough luminosity to drive the wind in the first place.

Simulations That Aged Well

For two decades, cosmological simulations including EAGLE, IllustrisTNG, and FABLE have hard-coded an early, vigorous active galactic nucleus feedback phase into their recipes. Without it, the simulated universe overproduces massive galaxies by factors that observers could no longer ignore. The recipe was a kludge: necessary to match the data, but with no direct observational anchor for how strong the early feedback actually had to be.

The new survey supplies that anchor. The measured kinetic-energy rates land squarely in the range simulators had assumed and could not previously verify. Dr Jan-Torge Schindler, who leads an Emmy Noether Group on early black holes at the Hamburg Observatory’s Quantum Universe cluster, contributed to identifying and characterising the parent quasar sample. His group’s earlier wide-field surveys provided the high-redshift candidates the JWST programme then followed up.

Outflows are driven in many directions by radiation pressure from the quasar’s extreme bright light, distinct from particle jets that carve a thin tunnel through a galaxy.

That description from Xiaohui Fan, a Regents Professor of astronomy at Arizona and the project’s principal investigator, captures the geometric difference that finally squares the observational record with the simulation playbook. Narrow jets were never going to expel gas across an entire galaxy disk. Wide radiation winds always could; until now nobody had caught enough of them in the act to claim the channel was common.

Stats Snapshot

• 22.2% of luminous early quasars show extreme galaxy-scale outflows, against 3.4% in the local universe.
• 8,400 km/s top wind velocity measured in the JWST sample.
• 46,000 solar masses per year peak mass outflow rate, enough to empty a Milky Way-mass gas reservoir in geological time.
• 100 million years typical active phase, comfortably long enough to lock in quenching.

The next test is straightforward. If quasar feedback is the dominant quenching channel at high redshift, the gas-poor early galaxies JWST has already catalogued should sit, statistically, near former or fading quasar hosts. The Nature paper sketches a follow-up plan using JWST’s MIRI instrument to look for cooler outflow phases at slightly later epochs, where the radiation pressure has eased but the galaxy has not yet recovered. The full arXiv preprint with the survey methodology lays out the comparison-sample selection criteria for groups planning to repeat the measurement with different oxygen tracers.

Frequently Asked Questions

What is a quasar?

A quasar is the visible signature of a supermassive black hole that is feeding actively on surrounding gas and dust. The infalling material forms a luminous accretion disk that can outshine every star in the host galaxy combined, which is why distant quasars look like point sources, or quasi-stellar objects, hence the name.

What does it mean to quench star formation?

Quenching describes the transition of a galaxy from actively forming new stars to producing few or none. It happens when the supply of cold molecular gas, the raw material for new stars, is either removed, heated, or prevented from collapsing into stellar nurseries. A quenched galaxy keeps its existing stars but stops growing.

How does the James Webb Space Telescope measure gas moving at thousands of kilometres a second?

The observatory’s NIRSpec instrument splits the light from a quasar’s host galaxy into a spectrum and looks at how much specific emission lines are shifted from their rest wavelengths. Gas moving towards Earth shifts to shorter (bluer) wavelengths; the size of the shift converts directly to a velocity using the Doppler formula.

Why are these outflows more common in the early universe?

Two ingredients compound at high redshift. Early galaxies hold denser gas reservoirs that couple efficiently to radiation pressure, and their central black holes were already remarkably massive, producing accretion luminosities high enough to launch wide-angle winds. Later in cosmic history, both ingredients weaken, so the extreme-outflow phase becomes much rarer.

What is the [O III] emission line and why is it the right tracer?

[O III] is a forbidden transition of doubly ionised oxygen at 5,007 angstroms. The transition only emits efficiently in low-density gas, the kind found in the body of a galaxy rather than near the black hole’s accretion disk. That property lets astronomers be sure the gas they are watching move is at kiloparsec scales, not just close to the central engine.

Does this discovery change the standard cosmological model?

No. It does the opposite: it confirms a long-standing prediction from cosmological simulations that needed a strong early active-galactic-nucleus feedback phase to match galaxy population statistics. The survey adds an observational anchor for a piece of the model that was previously calibrated by necessity rather than direct measurement.

Will future observations extend these results?

Yes. The same team has proposed follow-up programmes using JWST’s mid-infrared instrument to map cooler outflow phases at slightly later cosmic epochs. The European Extremely Large Telescope, due to begin science operations later this decade, will also be able to resolve outflow structure on smaller spatial scales than NIRSpec can reach.

Until follow-up surveys land, the dominant mechanism for shutting down galaxies in the first billion years of cosmic history has a name and a measured energy budget. The corpses match the weapon.