Yet the exact physical processes that couple AGN energy to the host galaxy are still debated. Feedback is commonly divided into two modes: ejective, or quasar-mode, feedback that expels gas through outflows, and preventative, or radio-mode, feedback that heats gas so it cannot cool and refuel star formation. Because these simulations contain many tunable parameters, models with very different feedback prescriptions can produce galaxies with roughly the same stellar properties, making direct observations essential.

Observationally, the challenge is timing. Most local quiescent galaxies quenched some eight to ten billion years ago, and the spectroscopic fingerprints of rapid quenching and powerful outflows fade within about a billion years. To identify the responsible physics, astronomers must catch galaxies in or shortly after the act of shutting down.

Post-starburst galaxies, identified by weak ionized gas emission and strong Balmer absorption lines from a dominant population of A-type stars, fit the bill: their spectra indicate a violent burst of star formation a few hundred million years before observation, followed by rapid quenching. Such galaxies account for less than one percent of massive galaxies locally but roughly five percent at redshift two, and as many as half of all galaxies entering the red sequence at that epoch may pass through a post-starburst phase. Before JWST, evidence linking outflows to quenching was largely circumstantial.

High-velocity outflows are found in about half of massive local post-starburst galaxies, and their strength declines with time since quenching, suggesting sudden quenching is driven by rapid ejection of cold gas. Deep ALMA observations showed that molecular gas reservoirs in nearby post-starburst systems decline faster than residual star formation alone can explain. At higher redshift, however, studies relied on small samples, stacked spectra or rest-frame color selections, and ionized gas tracers suggested that outflow rates in typical massive galaxies at cosmic noon fell below their star-formation rates, casting doubt on whether winds could meaningfully deplete the gas supply.

The key advance comes from sodium absorption. The Na I D doublet, at rest wavelengths of 5891 and 5897 angstroms, traces neutral gas because neutral sodium has a low ionization potential of 5.1 electronvolts and survives only in regions shielded by substantial columns of gas and dust. Unlike rest-frame ultraviolet absorption lines, which require a bright ultraviolet continuum and fail in dusty or quenching systems, Na I D is routinely detected in dusty, low-star-formation galaxies and is typically unsaturated, allowing robust constraints on optical depth and covering fraction.

Its main drawback is that at redshifts beyond about 0.7 the doublet shifts into the near-infrared, where atmospheric contamination made ground-based detection of the required stellar continuum impractically difficult, until JWST. The Blue Jay survey, a JWST Cycle 1 program targeting 150 galaxies at redshifts 1.7 to 3.5, provided the first comprehensive look at neutral outflows at cosmic noon, with findings since confirmed by the EXCELS, DeepDive and JADES surveys. Strong Na I D absorption was detected in roughly half of the massive galaxies, spanning star-formation rates from active to quiescent.

Notably, the absorption in redshift-two quiescent galaxies is an order of magnitude stronger than in quenched systems nearby and cannot be attributed to stellar atmospheres. Half of the detections are blueshifted by at least 100 kilometers per second, the signature of outflows, and blueshifted absorption in the Ca II H and K lines independently confirms the presence of cool neutral winds. The inferred neutral gas outflow rates are at least as large as the host galaxies' star-formation rates, corresponding to mass loading factors of one or more, with velocities of 200 to 1000 kilometers per second.

When both neutral and ionized outflows are measured in the same galaxies, the neutral phase dominates the outflowing mass but travels more slowly than the ionized gas, consistent with multiphase wind models. In quiescent galaxies with little ongoing star formation, several lines of evidence point to AGN as the engine: many show high-ionization emission lines characteristic of radiative AGN activity, ionized outflows reach 800 to 1000 kilometers per second, well above typical star-formation-driven winds, and in most cases supernovae cannot supply the energy or momentum the outflows carry, whereas the detected AGN can. Some outflows in quiescent galaxies may be fossils of past AGN episodes, since starburst-driven winds persist at most about 25 million years while post-burst ages are hundreds of millions of years, though low-luminosity AGN activity remains an alternative explanation.

The link to quenching is now direct in individual cases. Several massive quenched galaxies at redshifts one to four that shut down star formation within the last 300 million to one billion years show neutral outflows with mass loading factors of one to ten, sufficient to quench. Statistically, within the Blue Jay sample neutral outflows are detected in 75 percent of galaxies quenched within the last 500 million years, but in none of the more evolved systems, and ionized outflows appear exclusively in galaxies less than 500 million years past quenching.

At redshifts three to four, 23 to 30 percent of galaxies within 500 million years of quenching show neutral outflows, while older galaxies do not. These results establish ejective feedback as a compelling mechanism for the fast quenching pathway, estimated at 100 to 300 million years, as distinct from slow quenching over about a billion years attributed to gas starvation. What happens afterward is less clear.

About 71 percent of the redshift-two neutral outflows are too slow to escape their host halos, suggesting much of the ejected gas may eventually rain back in a galactic fountain. Indeed, JWST has revealed a strong neutral inflow feeding a massive galaxy at redshift 2.7 that has nevertheless been quiescent for a billion years, and weak systemic absorption in older quenched galaxies hints at lingering neutral gas. AGN heating may prevent rejuvenation: 20 to 50 percent of passive galaxies at redshifts one to five show emission line signatures of weak AGN activity or shocks, and radio jets appear to strengthen as quiescent galaxies age.

Episodic feedback, in which repeated cycles of inflow, AGN activity and outflow keep gas from forming stars, is another candidate for maintaining long-term quiescence. The new measurements also offer a way to test the feedback prescriptions embedded in cosmological simulations. The relation between stellar mass and outflow mass loading factor constrains where AGN feedback takes over from stellar feedback, and whether ejective or preventative modes dominate at high masses.

Because most prior measurements used ionized gas alone, and small multiphase samples suggest the cool component boosts loading factors by roughly an order of magnitude, JWST's neutral outflow census is poised to sharpen these comparisons considerably. Significant systematic uncertainties remain, however: converting sodium columns to hydrogen columns requires assumptions about sodium abundance, neutral fraction and dust depletion that are virtually unconstrained in winds, and outflow geometry is poorly known, with current size estimates ranging from one to 38 kiloparsecs. The assumed one-kiloparsec radius in several analyses makes published outflow rates conservative lower limits.

Extending this work to the earliest massive quiescent galaxies, where neutral outflows have now been detected at redshifts near four and even seven, may reveal how the first dead galaxies in the Universe met their end.

Subject of Research: Neutral gas outflows and rapid star-formation quenching in massive galaxies at cosmic noon observed with JWST

Article Title: Galactic outflows and star-formation quenching: a JWST perspective

Article References: Davies, R. L., & Belli, S. (2026). Galactic outflows and star-formation quenching: a JWST perspective. Astrophysics and Space Science, 371(10), Article 123. https://doi.org/10.1007/s10509-026-04655-4

Image Credits: AI Generated

DOI: 10.1007/s10509-026-04655-4

Keywords: JWST, galactic outflows, quenching, AGN feedback, post-starburst galaxies, cosmic noon, Na I D absorption, neutral gas, Blue Jay survey, mass loading factor, supermassive black holes, galaxy evolution