Fast-moving stars around an intermediate-mass black hole in ω Centauri

ω Centauri (ω Cen) is the Milky Way’s most massive globular cluster and is widely considered the stripped nucleus of an accreted dwarf galaxy. Its high mass, complex stellar populations, and proximity (D = 5.43 kpc) make it a prime target to search for an intermediate-mass black hole (IMBH). The research question is whether ω Cen hosts an IMBH, and if so, what constraints can be placed on its mass and position. Using an updated, deep proper-motion catalogue from more than 500 HST images spanning ~20 years, the authors identify a statistically significant overdensity of fast-moving stars in the cluster’s central 3 arcsec, with projected 2D velocities exceeding the cluster escape velocity if no IMBH is present. This motivates testing the IMBH hypothesis and deriving mass limits from stellar kinematics.

Multiple prior studies have reported evidence for an IMBH in ω Cen using integrated-light LOS velocity dispersion and dynamical modeling, with suggested masses around (3–4.7) × 10⁴ M⊙, but results have been controversial due to sensitivity to the assumed cluster center, possible degeneracy with centrally concentrated clusters of stellar-mass black holes, and velocity anisotropy effects. Anderson & van der Marel (2010) placed an upper limit of 1.2 × 10⁴ M⊙ using proper motions and a revised center, challenging earlier detections. Subsequent N-body comparisons have shown that both an IMBH and a dark cluster of stellar-mass black holes could reproduce some observations, and the absence of fast-moving stars in earlier PM catalogues favored the latter. Recent discoveries of a counter-rotating core in the central region underscore the complex kinematics and the importance of a well-determined center coincident with the AvdM10 center. The present study addresses a key missing piece noted in earlier work: the detection of fast-moving stars exceeding the escape velocity in the central region.

• Data: Archival HST imaging of ω Cen’s central region over >20 years, >500 images, leading to a high-precision proper-motion catalogue for 1.4 million stars with typical >20-year temporal baselines. Median PM errors in the center are ~6.6 μas yr⁻¹ (0.17 km s⁻¹) per component for bright stars.
• Sample selection: High-quality subset criteria to minimize spurious PMs: ≥20 astrometric measurements, temporal baseline ≥20 years, <15% rejected measurements (sigma clipping), PM errors <0.194 mas yr⁻¹ (~5 km s⁻¹) per component, reduced χ² < 10 for linear PM fits in RA and Dec, and location on the HST-based CMD main sequence (cluster membership). Uniform completeness within r ≲ 90″.
• Identification of fast-moving stars: Stars with total proper motion > 2.41 mas yr⁻¹ (escape velocity 62 km s⁻¹ at D=5.43 kpc) within 3″ of the AvdM10 photometric center were flagged. Seven stars (A–G) were found; for robust mass constraints the analysis focused on those with velocities at least 3σ above the escape speed (excluding B and G), after extensive astrometric quality checks (QFIT, radial excess, neighbor flux ratios) and multi-epoch image inspection.
• Contamination assessment: Foreground Milky Way contaminants estimated from the observed density at larger radii and the Besançon Galaxy model yield a number density 0.0026 arcsec⁻²; expectation value 0.074 contaminants within 3″. Poisson statistics show the chance of detecting five such stars by coincidence is P = 1.7 × 10⁻⁹; probability of ≥2 contaminants among five is P = 0.0026.
• Mass constraints from kinematics: Assuming the fast-moving stars are bound to a compact central object, the 2D velocities alone provide a lower limit on the IMBH mass. The minimal mass configuration depends on the unknown BH position; the position minimizing the mass lies ~0.3″ from the AvdM10 center. A combined constraint from five robust stars gives a firm lower limit of ~8,200 M⊙. Accelerations were then analyzed: individual accelerations are consistent with zero within 3σ (two >2σ), but joint modeling of velocities and acceleration limits via MCMC yields a stronger lower limit of 21,100 M⊙ at 99% confidence and prefers a BH position ~0.77″ northeast of the AvdM10 center.
• Model comparisons: The full PM velocity distribution within 10″ was compared to state-of-the-art N-body models with varying IMBH masses and alternatives (no IMBH; stellar-mass BH cluster). Models with no IMBH, a stellar-mass BH cluster, or very massive IMBHs (≳50,000 M⊙) are strongly disfavored; models with IMBH masses ~39,000–47,000 M⊙ best reproduce the fraction of fast-moving stars and overall distribution, though low-number statistics and some mismatches indicate a need for improved modeling.
• Additional considerations: The four innermost fast stars (rproj < 1″) are faint (M_F606W > 22.7) and preferentially on the blue side of the main sequence; the probability of all being this faint by chance is P = 0.013, hinting at possible capture or tidal interaction effects. The study also argues against explanations via stars bound to stellar-mass (≤100 M⊙) black holes or ejections from three-body interactions.

• Discovery of seven fast-moving stars within the central 3″ (0.08 pc) of ω Cen with 2D velocities exceeding the cluster’s escape velocity if no IMBH is present (v_esc = 62 km s⁻¹). The fastest, centremost star (A) has v_2D = 4.41 ± 0.08 mas yr⁻¹ = 113.0 ± 1.1 km s⁻¹, measured over 286 epochs across 20.6 years.
• Robust subset (≥3σ above escape speed) used for mass constraints after quality checks; excluding stars B and G has negligible impact on results.
• Foreground contamination rate is extremely low: 0.0026 arcsec⁻²; expected 0.074 stars within 3″; probability of seeing five contaminants by chance P = 1.7 × 10⁻⁹; probability of ≥2 contaminants among five P = 0.0026.
• Lower limit on central mass from 2D velocities alone: ≈8,200 M⊙, implying a compact massive object consistent with an IMBH. The minimum-mass BH position is ~0.3″ from the AvdM10 center (within its ±1″ uncertainty).
• Including acceleration constraints with MCMC increases the lower mass limit to 21,100 M⊙ at 99% confidence and favors an IMBH position 0.77″ NE of the AvdM10 center.
• N-body model comparisons of the inner 10″ PM distribution rule out no-IMBH, a stellar-mass BH cluster, and IMBH masses ≳50,000 M⊙, and favor IMBH masses around 39,000–47,000 M⊙; nonetheless, low-number statistics limit the precision of these inferences.
• The four innermost fast stars (rproj < 1″) are unusually faint (M_F606W > 22.7; P = 0.013 for random occurrence) and blue main-sequence members, potentially reflecting capture or tidal interaction processes near the IMBH.

The detection of multiple stars moving faster than the nominal escape velocity in the central 0.08 pc of ω Cen provides strong, direct dynamical evidence for a compact massive object that raises the local escape speed—consistent only with an IMBH rather than a concentration of stellar-mass black holes. The measured 2D velocities require a minimum central mass of ≈8.2 × 10³ M⊙, and joint velocity–acceleration analysis strengthens this to ≥2.11 × 10⁴ M⊙ (99% confidence). Comparisons to N-body models indicate that an IMBH of order 4 × 10⁴–4.7 × 10⁴ M⊙ best matches the observed inner velocity distribution and fraction of fast movers, while alternatives (no IMBH or a stellar-mass BH cluster) are disfavored. These results address the long-standing controversy over an IMBH in ω Cen by providing the previously lacking evidence of fast-moving, likely bound stars analogous to the S-stars around Sgr A*. The inferred IMBH has important implications for black hole demographics in low-mass galaxies and star clusters, adding a nearby benchmark system where the dynamics of multiple bound stars can be studied in detail. Remaining uncertainties in the precise IMBH mass and position, as well as line-of-sight distances and velocities of the stars, motivate further targeted observations and improved dynamical modelling.

This study identifies seven fast-moving main-sequence stars in the central 3″ of ω Centauri whose velocities exceed the cluster’s escape speed in the absence of a massive central object. Their presence implies a compact massive object consistent with an intermediate-mass black hole. Kinematic constraints from 2D velocities establish a firm lower mass limit of ~8,200 M⊙, strengthened to ≥21,100 M⊙ (99% confidence) when including acceleration limits; comparisons to N-body models favor IMBH masses ~39,000–47,000 M⊙. The findings provide one of the strongest cases for an IMBH in the local universe and a unique nearby laboratory—second only to Sgr A*—to study multiple bound stellar orbits. Future work should include: (i) dynamical modeling that jointly fits PMs, LOS velocities, and the effects of mass-segregated dark remnants; (ii) spectroscopic IFU observations (e.g., VLT MUSE, JWST NIRSpec IFU) to obtain LOS velocities; and (iii) deeper, higher-precision astrometry (e.g., JWST NIRCam, VLTI GRAVITY, ELT MICADO, VLT MAVIS) to detect accelerations and additional tightly bound stars for direct IMBH mass measurements. Extending similar searches to other likely accreted nuclear clusters (e.g., M54) and to fainter stars in other globular clusters is also motivated.

• Unknown line-of-sight distances and velocities for the fast-moving stars; constraints rely on projected 2D motions and assume bound orbits.
• The exact position and mass of the IMBH remain uncertain; accelerations are weak and largely consistent with zero at present precision.
• Low-number statistics of fast-moving stars limit comparisons to models and the robustness of inferred mass ranges; some mismatches with the overall velocity distribution indicate the need for improved modeling.
• Sensitivity to the adopted cluster center, though the minimum-mass position is consistent with the AvdM10 center within its uncertainty.
• Potential selection effects in previous catalogs (now mitigated by depth/coverage) and residual contamination, although the expected contamination rate in the central 3″ is very low.
• Current non-detections at X-ray/radio imply an extremely low accretion rate, offering limited independent constraints on BH properties.