JWST Survey Finds Slow, Wide Molecular Hydrogen Winds Are Common Around Planet-Forming Disks

Somewhere in the dusty disk around a young star, a planet is trying to gather gas. It may have only a few million years before the disk's gas disperses or is locked into worlds. What drives that clearing is not fully settled. A new JWST survey now points to a common candidate: faint, slow, wide-angle winds of molecular hydrogen, traced by warm H2 emission extending away from the disks.

A survey of 34 disks in the infrared light of H2

Molecular hydrogen, H2, emits infrared light when its molecules rotate. JWST's Mid-Infrared Instrument (MIRI) and its Medium Resolution Spectrometer can detect several of those pure rotational lines, a direct signature of warm gas. The new work draws on the JWST Disk Infrared Spectroscopic Chemistry Survey (JDISCS), which observed 30 protoplanetary disks, plus four archival disks from the Mid-INfrared Disk Survey (MINDS). In total, the authors analyzed 34 young systems.

To isolate faint extended emission, the team subtracted the unresolved central source and compact disk line emission from each line map. The H2 maps are shown over ALMA 240 GHz continuum images of the same disks.

Cones, rings, and bubbles—and slow, wide, warm gas

Extended H2 emission turned out to be common. Sixteen of the 34 disks show clear morphological signs of a wind. In disks tilted to our line of sight, the H2 maps reveal conical structures: some one-sided (monopolar), some two-sided (bipolar). In disks seen nearly face-on, the emission often forms rings or bubbles around the star. Both patterns match what a wide-angle wind should look like from different viewing angles.

The team reports that after excluding four higher-mass Herbig Ae/Be stars, which showed no H2 emission, over 50 percent of the remaining sample shows clear extended H2 emission attributable to winds. If the four known binaries are excluded too, that fraction rises to 61 percent. Not every disk has an identical wind: several sources are faint, patchy, or contaminated by foreground or background H2 from nearby molecular clouds.

For ten disks, the researchers modeled the wind shape by fitting a power law to its edges. The median half-opening angle is 45 degrees, much wider than a narrow jet. The gas is warm, with a median rotational temperature of about 624 kelvin and a median column density of about 10^18.6 molecules per square centimeter. The velocities are modest: a median of 4.2 kilometers per second, with large uncertainties. For comparison, protostellar outflows typically reach about 10 to 30 kilometers per second.

Those properties, the authors write, are consistent with slow magnetohydrodynamic (MHD) winds—outflows driven by magnetic fields that grip the rotating disk and fling gas outward. Photoevaporation, in which high-energy radiation heats and ionizes the disk surface until gas escapes, remains a possible alternative. The data cannot yet choose definitively between them.

How much gas can a molecular wind carry—and what remains uncertain

The survey estimates a median mass-loss rate of about 10^-9 solar masses per year. If molecular winds are the dominant mechanism clearing disks, that rate would empty a disk of 2 to 3 Jupiter masses in roughly 2 to 3 million years. That is striking because current estimates for typical protoplanetary disk lifetimes are about the same.

The wind rate is not simply an echo of accretion. Across the sample, wind mass-loss rates span about a factor of 100, while accretion rates onto the stars span about a factor of 10,000. The average ratio is roughly 0.1, but the two do not correlate strongly. Accretion traces near-instantaneous conditions; the H2 wind averages over a dynamical time of a few hundred years. They may be different clocks.

Several caveats remain. The study is a preprint, not a final peer-reviewed result. Its H2 maps trace only warm gas at about 650 kelvin. A colder or hotter component could hide from MIRI, and the derived molecular mass-loss rate is a lower limit on total mass loss because the wind's molecular fraction is unknown. The largest uncertainty is the dynamical timescale used to convert wind mass into a mass-loss rate. Point-spread-function subtraction can leave artifacts, and some disks sit in front of bright background H2 from nearby molecular clouds. The authors also note that only ten disks were modeled in detail, and that more detailed H2 predictions, especially for MHD winds, are needed before photoevaporation and MHD winds can be definitively distinguished.

Even with those caveats, the picture is larger than a few odd outflows. JWST is showing that many planet-forming disks exhale slowly, widely, and in infrared light from molecular hydrogen. Those winds may help set the gas deadline for forming giant planets, removing gas that would otherwise feed growing worlds. The next step is to turn that deadline into a detailed clock—one that connects magnetic fields, starlight, and the quiet molecular breath of a young planetary system.