Origin of the initial mass function
Why do newly formed stars seem to follow the same distribution of masses regardless of where they form?
Summary
Why it is open
The initial mass function (IMF) is the distribution of stellar masses at birth [12]. Because a star's properties are decided primarily by its mass, Patrick Hennebelle and Michael Grudić call the IMF critical to the structuring of the Universe [12], and estimates of the masses and star formation rates of distant galaxies must assume one [10, 14].
Two questions are mixed together. One is physical: what sets the power-law tail at high masses and the peak at low masses [12]. The other is empirical: whether the IMF is the same everywhere [5, 12]. The IMF is not directly observable, Hennebelle and Grudić explain: systems old enough for their low-mass stars to reach the main sequence have already lost their massive stars, while younger systems need model assumptions, so every measurement depends on models [12].
On universality, reviews differ in emphasis. In 2010 Nate Bastian, Kevin Covey and Michael Meyer found "no clear evidence that the IMF varies strongly and systematically as a function of initial conditions after the first few generations of stars" [5]. In 2024 Hennebelle and Grudić wrote that "the strong hypothesis of a true IMF universality is unlikely", while noting that the variations measured in nearby galaxies are typically not large [12]. For distant and massive galaxies, whose stars cannot be counted one by one, different methods can disagree about the same kind of galaxy [8, 12].
No body sets a bar for a resolution. Hennebelle and Grudić list what is still needed: linking the masses of dense gas cores to the stars they form, simulations whose results do not depend on numerical resolution, and complete models of well-observed regions [12].
What is known
Standard forms (1955–2003). In 1955 Edwin Salpeter derived an "original mass function" from the luminosity function of stars near the Sun, assuming among other things that stars had formed at a uniform rate for five billion years [1]. Where the logarithm of mass in solar units lies between −0.4 and +1.0, the number of stars per logarithmic mass interval fell as mass to the power −1.35; whether a steeper drop above 10 solar masses was real was "not yet clear", he wrote [1]. In 2001 Pavel Kroupa summarised the constraints as a power law with index 2.3, his convention's Salpeter value, above 0.5 solar masses, flattening to 1.3 between 0.08 and 0.5 and to 0.3 between 0.01 and 0.08 [2]. In 2003 Gilles Chabrier described the IMF as a power law above about one solar mass and a lognormal form below [3]. The high-mass slope stays close to Salpeter's, and the peak usually lies in a narrow range of about 0.1–0.3 solar masses [12].
A common IMF (2001–2026). Studies of the field, young clusters and associations, and old globular clusters suggest, Bastian and colleagues concluded, that the vast majority were drawn from a "universal" IMF [5]. Kroupa found that Poisson noise and the dynamical evolution of star clusters reproduce "quite well" the observed scatter in measured slopes [2]. In the Andromeda galaxy, a homogeneous sample of young clusters gave in 2015 a slope above about 2 solar masses with intrinsic scatter consistent with zero [12]. In September 2026 Keyi Ding and colleagues counted more than 10,000 stars down to about 0.15 solar masses in the ultrafaint dwarf galaxy Boötes I, whose old stars are very poor in metals, with a logarithmic iron-to-hydrogen ratio about 2.4 below the Sun's, and found IMF parameters consistent with the Milky Way's [15].
Bottom-heavy IMFs in giant ellipticals (2010–2026). Pieter van Dokkum and Charlie Conroy detected, in luminous elliptical galaxies, sodium and iron hydride features that are strong only in stars below 0.3 solar masses. They inferred that such stars make up over 80% of the stars and over 60% of the stellar mass, and an IMF probably slightly steeper than Salpeter's between 0.1 and 1 solar mass [6]. From dynamical models of the ATLAS3D sample, Michele Cappellari and colleagues reported in 2012 a systematic variation of the IMF with stellar mass-to-light ratio, "producing differences up to a factor of three in mass" [7]. Gravitational lensing of three nearby ellipticals with velocity dispersions above 300 km/s gave Russell Smith, John Lucey and Conroy mass-to-light ratios consistent with a Milky Way IMF; they firmly rejected very heavy IMFs and called their results "apparently in conflict" with spectroscopic bottom-heavy IMFs [8]. The trend towards heavier IMFs in the centres of more massive galaxies persists in later work [12]. In August 2026 Chloe Cheng and colleagues reported in Nature Astronomy, from JWST spectra of nine massive quiescent galaxies at redshift about 0.7, that the most massive have excess low-mass stars and that the oldest, formed before redshift 5, has the most bottom-heavy IMF; if JWST's early massive galaxies were similar, their masses would rise by a factor of about 4 ± 1 [14].
Top-heavy IMFs (2018–2025). From spectroscopy of 247 stars above 15 solar masses in 30 Doradus, a starburst region in the Large Magellanic Cloud, Fabian Schneider and colleagues found 32 ± 12% more stars above 30 solar masses than a Salpeter IMF predicts [9]. Several dense Local Group clusters and the Galactic nuclear cluster also show relatively top-heavy IMFs, although the dynamical evolution of dense clusters can bias such measurements in either direction [12]. Using the ratio of the isotopologues ¹³CO and C¹⁸O, which dust does not affect, Zhi-Yu Zhang and colleagues reported a top-heavy IMF in all four dusty starbursts they observed at redshifts about 2–3 [10]. For the early universe, Anne Hutter and colleagues showed in 2025 that a model in which the IMF becomes more top-heavy at high gas density reproduces the observed ultraviolet luminosity functions of galaxies at redshifts 5–15; this is a model result, not a measurement [13].
Metallicity and age (2001–2023). Kroupa suggested in 2001 that higher-metallicity environments produce relatively more low-mass stars [2]. From about 93,000 M dwarfs within 100–300 parsecs, Jiadong Li and colleagues reported in 2023 that early-formed populations contain fewer low-mass stars than the canonical IMF, and that in present-day populations the share of low-mass stars rises with metallicity [11].
Main approaches
Competitive accretion. Stars in a forming cluster compete for gas, and more massive stars attract more of it, so they grow further; the idea was proposed by Zinnecker in 1982 and developed by Bonnell and colleagues in 2001 [12]. Strongest evidence: the model predicts a power law, and simulations that place accreting particles in a molecular cloud develop one [12]. Main difficulty: the predicted slopes are close to, but slightly shallower than, Salpeter's [12].
Gravo-turbulent fragmentation. Supersonic turbulence creates density fluctuations, and those massive enough to collapse become the gas reservoirs of future stars [4, 12]. Strongest evidence: the 2008 analytic theory of Hennebelle and Chabrier gives a power law with a lognormal cut-off and recovers the Salpeter slope for the turbulent velocity spectrum found in observations and simulations [4]; dense cores have mass functions similar in shape to the IMF [12]. Main difficulty: the predicted peak depends on gas density, temperature and turbulent speed, which vary between clouds; core mass functions in the W43 region are flatter than the IMF, and the inferred core peak shifts with the resolution of the observations [12].
Small-scale physics sets the peak. The characteristic mass comes from physics that depends only weakly on environment: heating of the surrounding gas by accreting protostars, proposed by Bate, Krumholz and others, or the first hydrostatic core, a gas object of about 0.03 solar masses that must form before a protostar can and whose tidal pull suppresses fragmentation nearby, proposed by Hennebelle and colleagues [12]. Strongest evidence: both predict masses near the observed peak with weak dependence on density; the tidal model gives about ten times the core mass, around 0.3 solar masses [12]. Main difficulty: simulations of protostellar heating have reached conflicting results, and the computed peak can depend drastically on resolution and on how simulations represent forming stars [12].
Open sub-questions
- Lenses against spectra. In nearby lens ellipticals with velocity dispersions above 300 km/s, do sodium and iron hydride features measured in the same galaxies agree with the Milky Way-like mass-to-light ratios from lensing [6, 8]?
- The oldest quiescent galaxies. Do other quiescent galaxies that formed before redshift 5 show IMFs as bottom-heavy as the oldest galaxy studied by Cheng and colleagues [14]?
- Dynamics in 30 Doradus. How much of the excess of massive stars in 30 Doradus could mass segregation produce, and how much would ejection of massive stars hide [9, 12]?
- Metal-poor dwarfs. Do resolved counts in other equally metal-poor ultrafaint dwarfs find a Milky Way-like low-mass IMF, as in Boötes I, or the deficit of low-mass stars reported for early-formed populations near the Sun [11, 15]?
- From cores to stars. In regions with flat core mass functions such as W43, does the stellar mass function that forms share the flatter slope [12]?
- A converged peak. Does the simulated IMF peak stay fixed when the first hydrostatic core is resolved and the treatment of forming stars is varied [12]?
Work on this in Papex
If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.
Key references
- Salpeter, E. E. (1955). The luminosity function and stellar evolution. The Astrophysical Journal 121, 161–167. 10.1086/145971 · Checked 15 Oct 2026
- Kroupa, P. (2001). On the variation of the initial mass function. Monthly Notices of the Royal Astronomical Society 322(2), 231–246. 10.1046/j.1365-8711.2001.04022.x · Checked 15 Oct 2026
- Chabrier, G. (2003). Galactic stellar and substellar initial mass function. Publications of the Astronomical Society of the Pacific 115(809), 763–795. 10.1086/376392 · Checked 15 Oct 2026
- Hennebelle, P. & Chabrier, G. (2008). Analytical theory for the initial mass function: CO clumps and prestellar cores. The Astrophysical Journal 684(1), 395–410. 10.1086/589916 · Checked 15 Oct 2026
- Bastian, N., Covey, K. R. & Meyer, M. R. (2010). A universal stellar initial mass function? A critical look at variations. Annual Review of Astronomy and Astrophysics 48, 339–389. 10.1146/annurev-astro-082708-101642 · Checked 15 Oct 2026
- van Dokkum, P. G. & Conroy, C. (2010). A substantial population of low-mass stars in luminous elliptical galaxies. Nature 468(7326), 940–942. 10.1038/nature09578 · Checked 15 Oct 2026
- Cappellari, M., McDermid, R. M., Alatalo, K., Blitz, L., Bois, M., Bournaud, F. et al. (2012). Systematic variation of the stellar initial mass function in early-type galaxies. Nature 484(7395), 485–488. 10.1038/nature10972 · Checked 15 Oct 2026
- Smith, R. J., Lucey, J. R. & Conroy, C. (2015). The SINFONI Nearby Elliptical Lens Locator Survey: discovery of two new low-redshift strong lenses and implications for the initial mass function in giant early-type galaxies. Monthly Notices of the Royal Astronomical Society 449(4), 3441–3457. 10.1093/mnras/stv518 · Checked 15 Oct 2026
- Schneider, F. R. N., Sana, H., Evans, C. J., Bestenlehner, J. M., Castro, N., Fossati, L. et al. (2018). An excess of massive stars in the local 30 Doradus starburst. Science 359(6371), 69–71. 10.1126/science.aan0106 · Checked 15 Oct 2026
- Zhang, Z.-Y., Romano, D., Ivison, R. J., Papadopoulos, P. P. & Matteucci, F. (2018). Stellar populations dominated by massive stars in dusty starburst galaxies across cosmic time. Nature 558(7709), 260–263. 10.1038/s41586-018-0196-x · Checked 15 Oct 2026
- Li, J., Liu, C., Zhang, Z.-Y., Tian, H., Fu, X., Li, J. & Yan, Z.-Q. (2023). Stellar initial mass function varies with metallicity and time. Nature 613(7944), 460–462. 10.1038/s41586-022-05488-1 · Checked 15 Oct 2026
- Hennebelle, P. & Grudić, M. Y. (2024). The physical origin of the stellar initial mass function. Annual Review of Astronomy and Astrophysics 62, 63–111. 10.1146/annurev-astro-052622-031748 · Checked 15 Oct 2026
- Hutter, A., Cueto, E. R., Dayal, P., Gottlöber, S., Trebitsch, M. & Yepes, G. (2025). Astraeus X. Indications of a top-heavy initial mass function in highly star-forming galaxies from JWST observations at z > 10. Astronomy & Astrophysics 694, A254. 10.1051/0004-6361/202452460 · Checked 15 Oct 2026
- Cheng, C. M., Slob, M., Kriek, M., Beverage, A. G., van Dokkum, P. G., Bezanson, R. et al. (2026). Hidden mass in early galaxies revealed by bottom-heavy initial mass functions. Nature Astronomy, published online 18 August 2026 (volume not yet assigned on 2026-10-15). 10.1038/s41550-026-02932-4 · Checked 15 Oct 2026
- Ding, K., Gennaro, M., Avila, R. J., Ricotti, M., Beaton, R. L., Boyer, M. L. et al. (2026). Probing the initial mass function in the early universe: direct measurements in the Boötes I ultrafaint dwarf galaxy with JWST/NIRCam. The Astrophysical Journal 1009(2), 94. 10.3847/1538-4357/ae9f3e · Checked 15 Oct 2026
Source
Open, as listed by Wikipedia · Unsolved problems in astronomy. Checked 08 Oct 2026.
The status is the source’s. Papex does not decide whether a problem is solved.
Records on this problem
No public record addresses this problem yet.