A paper accepted for publication in The Astrophysical Journal proposes that an intermediate-mass black hole of roughly 10,000 solar masses orbits the supermassive black hole at the center of the Milky Way, and that this hidden companion may solve a longstanding puzzle about the Galactic Center’s young star populations.
The authors, led by Xiaochen Zheng of the Beijing Planetarium, used N-body simulations to test whether an intermediate-mass companion (IMC) on a steeply inclined orbit could produce the observed distribution of three distinct groups of young stars within the central half-parsec of the galaxy. The answer, they found, is that it can, through a set of three gravitational interactions that each sculpt one population.
The problem
Within about 0.5 parsecs (1.6 light-years) of Sagittarius A*, the Milky Way’s central supermassive black hole, there are three populations of young stars, all with short lifetimes of 6 to 15 million years, meaning they formed recently and in roughly the same event:
1. The S-stars, about 40 B-type stars on highly eccentric, randomly oriented orbits squeezed into the inner 0.04 parsecs
2. The clockwise disk stars, O-type and Wolf-Rayet stars in an orderly disk from 0.04 to 0.5 parsecs
3. The off-disk stars, the majority, on highly inclined and eccentric orbits, including some that are counter-rotating
Explaining all three populations from a single formation event, a gas disk that fragmented into stars approximately 6 million years ago, has been difficult. The clockwise disk is natural enough, but the S-stars should not have such random orientations, and the off-disk stars should not exist at all if the gas disk was the only gravitational influence.
The proposed solution
The Zheng team’s simulations place an intermediate-mass companion of 10,000 solar masses on an orbit at 0.35 parsecs from Sgr A*, with an eccentricity of 0.3 and an inclination of 120 degrees, steeply tilted relative to the original gas disk. The companion does three things:
- Off-disk stars (the outermost population): The tilted companion’s gravitational pull drives von Zeipel-Lidov-Kozai (vZLK) oscillations, cycles that stretch and tilt orbits over millions of years, creating the observed population of highly inclined and counter-rotating stars.
- Clockwise disk stars (the mid-layer): As the gas disk depleted, a sweeping secular resonance moved inward from 0.2 parsecs to 0.04 parsecs, moderately exciting eccentricities without destroying the disk structure.
- S-stars (innermost): High-eccentricity stars injected by the companion scatter off inner stars through resonant relaxation, randomizing their orbits and creating a zone of avoidance in their eccentricity-pericenter distribution that matches observations.
A candidate already exists
The researchers identify IRS-13E, a known infrared source approximately 0.13 parsecs in projected separation from Sgr A*, as the physical location of the companion. IRS-13E has been debated for years: some observations suggest it is a genuine star cluster, while others argue it could be a chance alignment of unrelated stars. The new model provides a dynamical reason for it to be a real, bound object, and one massive enough to be the IMC.
Caveats
The paper is careful to call it an “intermediate-mass companion” rather than definitively an intermediate-mass black hole. The companion could be a compact stellar cluster instead. As Albert Zijlstra of the University of Manchester told New Scientist, “All the potential ones that have been found so far in this mass range have fallen through, mostly due to lack of evidence.” Existing GRAVITY interferometry observations place stringent upper limits on the mass of any central companion, and the X-ray emission from IRS-13E could be from Wolf-Rayet wind collisions rather than black hole accretion. The model shows that this scenario is possible, not that it is the actual explanation. Confirmation will require extended observations of IRS-13E’s proper motion over time.
Source: Zheng, X., Wang, L., Lin, D.N.C. et al. (2026). Accepted for publication in The Astrophysical Journal. arXiv:2606.08971. DOI: 10.3847/1538-4357/ae71ba