Gaia Data Show That the Initial Mass Function Is Not Universal

Do all star-forming regions birth stars with the same distribution of masses? Theory suggests that they do not, but researchers often assume they do for simplicity. Now, ultra-precise observations from Gaia have demonstrated that the initial mass function varies across star clusters in the Milky Way.

A Universal Initial Mass Function?

When a molecular cloud fragments, collapses, and births a star cluster, how many stars form, and what are their masses? The answer to that question is what astronomers term the initial mass function, or IMF: the distribution of masses of stars born in a cluster.

The IMF is important in many areas of astronomy, and it’s relied on especially heavily in studies of distant galaxies. The light from these galaxies is dominated by high-mass stars, with the emission from lower-mass stars lost in the glare, so astronomers must use the IMF to reconstruct the stellar populations of these far-off locales.

Though theory suggests that the IMF should vary with factors like temperature and metallicity, demonstrating these variations has been difficult, even within our own galaxy. Because of this, researchers often adopt a universal IMF drawn from observations of star clusters in the Milky Way. But is the IMF truly universal?

The More the Slope Changes, the More the Break Mass Stays the Same

Determining whether the IMF is universal is more complicated than simply comparing the stars in one cluster to those in another. That’s because what astronomers measure in present-day star clusters is not the initial mass function, but a version of the IMF that has been skewed by millions to billions of years of evolution; over time, high-mass stars die and low-mass stars get ejected, warping the shape of the mass function.

However, as demonstrated in a recent research article led by Charles L. Steinhardt (University of Missouri), certain aspects of the IMF may remain unchanged as clusters evolve. When the mass function of a star cluster is described using a broken power law (commonly called a “Kroupa-like” IMF in reference to work by Pavel Kroupa), the slopes of the power-law segments are affected by stellar evolution and dynamical interactions. Crucially, these effects appear not to alter the break mass — the mass at which the IMF transitions from one slope to another. Therefore, differences in the break mass from cluster to cluster reflect differences in the underlying IMF rather than the stellar and dynamical evolution of the clusters.

Insights from Gaia

Thanks to the Gaia spacecraft, which made precise observations of more than 2 billion stars in the Milky Way and beyond, it’s now possible to measure and compare the break masses of individual star clusters. Out of an initial catalog of 7,167 Milky Way clusters observed with Gaia, Steinhardt’s team selected just 110 that were of high enough quality and contained a sufficient number of stars to measure cluster break masses accurately.

The team found that the break mass differs considerably between the clusters in their sample, showing that the IMF is not universal. The observed variety of break masses appears to be linked to the age of the cluster, but instead of reflecting the impact of stellar and dynamical evolution — which have no effect on the break mass — these differences reflect the properties of the clouds in which the clusters were born. This implies that the properties of star-forming clouds in the Milky Way have, on average, changed over cosmic time. Thus, a star cluster that formed billions of years ago would have had a different IMF than one forming today.

This finding has implications for studies of high-redshift galaxies, where the properties of star-forming clouds are markedly different from those in the Milky Way. Going forward, Steinhardt and coauthors recommend developing a flexible parameterization of the IMF to capture how this critical function behaves in environments vastly different from our own.

Citation

“Direct Evidence for Stellar Initial Mass Function Variation in the Milky Way,” Charles L. Steinhardt et al 2026 ApJL 1005 L40. doi:10.3847/2041-8213/ae7444