Stellar motions support an excess of small stars in massive galaxy cores

A galaxy’s brightness does not reveal its stellar mass on its own. Converting light into mass requires an assumption about the mix of stars that formed there. Numerous faint, low-mass stars can add substantial mass without contributing much light.

A preprint led by Martje Slob uses James Webb Space Telescope observations to test that assumption in nine massive, quiescent galaxies—systems with little ongoing star formation—at redshift around 0.7. The researchers find that stellar motions support an excess of low-mass stars in the cores of the most massive systems, provided that this excess does not extend unchanged throughout each galaxy.

The study offers a dynamical check on earlier spectroscopic results, not a direct measurement of how the stellar mix changes from centre to outskirts.

Two ways to estimate mass

The relevant stellar mix is called the initial mass function, or IMF: the distribution of masses with which stars are born. Many galaxy studies adopt an IMF calibrated to the Milky Way. A “bottom-heavy” IMF instead contains a greater proportion of low-mass stars.

Earlier work by Chloe M. Cheng and colleagues examined absorption features in these galaxies’ combined starlight that are sensitive to dwarf stars. Seven galaxies overlap with the new dynamical sample. Three favour bottom-heavy IMFs, implying stellar masses roughly two to four times those inferred with a Milky Way-like IMF.

The new paper asks whether those inferred stellar masses fit within the mass allowed by gravity. Stellar motions respond to the gravitational pull of both visible and unseen matter. If an estimated stellar mass exceeds the dynamical mass, something in the estimates or their assumptions must be wrong.

Modelling what Webb actually sees

The JWST-IMFERNO programme obtained 31.2 hours of on-source spectroscopy with NIRSpec’s microshutter array. Nine of its thirteen targets provided enough reliable, spatially resolved measurements for this analysis. NIRCam images supplied the galaxies’ light distributions.

The researchers measured stellar velocities and the spread of velocities along each spectroscopic slit, using absorption features around the calcium triplet. That spread, called velocity dispersion, helps constrain the gravitational mass even where rotation is difficult to recover.

A slit samples only one spatial direction through a galaxy. Its placement and angle, together with instrumental blurring and light lost outside the slit, affect the observed motions. The team therefore built three-dimensional dynamical models using Jeans Anisotropic Modelling, then simulated how NIRSpec would observe them. Comparing those simulated observations with the data yielded mass estimates that account for these effects.

Three assumptions, three mass budgets

With a Milky Way-like IMF, every galaxy has less inferred stellar mass than dynamical mass: a physically possible result. But, assuming negligible gas, the remaining mass implies an average dark matter fraction within one effective radius of 54%, with a quoted uncertainty of 6 percentage points. The effective radius encloses half the galaxy’s light. The authors regard this fraction as unusually high compared with nearby early-type galaxies and simulation expectations.

Applying the spectroscopically measured IMF uniformly across each of the seven overlapping galaxies moves the estimates in the other direction. The mean inferred central dark matter fraction becomes 2%, with an uncertainty of 18 percentage points. In two galaxies, the stellar mass exceeds the dynamical estimate at about the two-sigma level.

The third scenario applies the measured IMF correction only inside one effective radius, retaining a Milky Way-like IMF outside. This produces consistent stellar and dynamical masses and a mean inferred central dark matter fraction of 28%, with an uncertainty of 12 percentage points. The authors consider this a more plausible mass budget.

Crucially, that transition is an imposed step function, not an observed boundary. It illustrates a possible radial change in the IMF; it does not establish its actual shape.

What remains uncertain

The sample is small, and all its galaxies occupy an overdense environment. The findings therefore do not establish a universal stellar mix for quiescent galaxies.

The dynamical models also assume axisymmetry. Their additional mass component cannot cleanly distinguish dark matter from changes in stellar mass-to-light ratio. The quoted dark matter fractions are conditional mass-budget estimates, not direct measurements of the dark matter halo.

Another assumption enters when the researchers estimate total dynamical mass by doubling the mass enclosed within one effective radius. If mass and light have different radial distributions—as an IMF gradient could cause—the half-mass and half-light radii need not coincide.

Within those limits, the results support a picture in which low-mass-star-rich cores were already present at redshift 0.7, with subsequent growth occurring mainly in galaxy outskirts through minor mergers. That history is consistent with the measurements, not uniquely demonstrated by them. Direct measurements of IMF changes across these galaxies, and observations of larger samples, are needed to test it.