A preprint reports that gas from AGB (asymptotic giant branch) stars gathered into a growing, disk-like structure around a central intermediate-mass black hole (IMBH) in a fiducial simulation. Gas inside 0.1 parsec of the black hole reached 481.7 solar masses after 0.188 million years. The calculation stopped short of modeling IMBH growth, so the result is a simulated gas reservoir, not mass shown to have been added to the black hole.
Inside the model
The study used smoothed-particle hydrodynamics, or SPH, simulations that follow AGB gas as particles to ask which conditions are needed to form massive disks around an IMBH. The model grid varied UCD stellar mass, black-hole mass, the number of AGB stars, gas temperature and wind velocity. Each IMBH was represented as a collisionless particle; black-hole growth and feedback on the interstellar medium were not modeled.
Across the model grid, disk mass increased more with UCD mass than with IMBH mass. For 100-solar-mass IMBHs, a majority of the gas present could be trapped in a disk. The result describes how these modeled systems behaved, not direct evidence that black holes in observed UCDs grow through the process.
What changed the disk
Under sufficiently favorable combinations of AGB-star numbers and UCD mass, the study’s conclusion says approximately 60% of the total available AGB gas could be trapped by a central black hole. That is a model-dependent capture estimate, and the simulations did not resolve the subsequent inflow into the black hole.
The results also differed with the assumed initial mass function, or IMF, the mix of stellar masses in the model. Top-heavy IMF models were associated with higher-mass disks over time than standard-IMF models. The study does not establish which IMF describes real UCDs; it shows how the modeled disk outcome varied with the assumed AGB supply.
Star formation was associated with gas removal and gravitational disk disruption in the simulations. Removing star formation yielded larger, longer-lived disks. This comparison is a contrast between model setups, so it does not establish a causal effect in real UCDs.
Black-hole mass fractions tracked a separate pattern. Higher fractions were associated with earlier, burst-like star-formation histories, while lower fractions had star-formation-rate profiles more like constant star formation. The reported relationship is model-specific.
The position comparison produced a marked contrast. With the black hole at 0.1 parsec, the modeled gas reservoir reached about 1,750 solar masses. With it at 8.0 parsecs, the reservoir reached about 600 solar masses and retained gas for about 0.4 million years. The more central placement was associated with a larger and longer-lived central gas reservoir.
From simulations to observed systems
To connect the model with observed compact systems, the researchers compared it with UCD data from Mieske et al. (2013), focusing on systems with high predicted black-hole mass fractions. For the initial-black-hole calculation, they chose a 42% radiative efficiency, an IMF slope of 1.4 and an available AGB mass equal to 24.3% of UCD mass.
Using the stated typical IMBH upper-limit criterion, 12 of the 17 presented UCD entries met it. Some inferred minimum initial masses were negative. Those numbers come from an extrapolation based on inferred black-hole masses and the selected capture, IMF and radiative-efficiency assumptions, not from direct measurements of black-hole growth.
A clue observers could test
The simulations also produced an exploratory kinematic clue. One modeled case contained four to six stars classified as high velocity, while only the high-mass IMBH models consistently produced a non-zero high-velocity count. The result is a pattern for observers to compare with data, not a confirmed signature.
Taken together, the calculations describe a possible route in which AGB gas gathers around an IMBH in sufficiently massive UCD models, with central black-hole placement linked to greater gas retention. But the crucial final step - gas inflow and conversion of disk material into black-hole mass - was not simulated. The proposed path from an IMBH to a massive black hole therefore remains a model-based extrapolation whose estimates rest on choices about the IMF, capture fraction and radiative efficiency.
Paper data and sources
Original title: Formation of Accretion Disks around Intermediate-Mass Black Holes in Ultra Compact Dwarf Galaxies and Nuclear Stellar Clusters
Authors: Sebastian Darr, Kenji Bekki
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text