A quarterly-updated synthesis of the multi-channel observational evidence for an intermediate-mass black hole in NGC 5139 (Omega Centauri). Each channel's constraints, internal tensions, and the combined posterior are reviewed in turn.
The most direct mass constraint comes from stellar kinematics (Häberle et al. 2024, Nature): seven stars within 0.08 pc of the cluster center with velocities implying a compact dark mass of MIMBH ≥ 8,200 M☉, a firm lower bound rather than a point estimate. This constitutes the best available positive evidence for an IMBH in Omega Centauri.
The strongest upper limit is from millisecond pulsar timing alone (TRAPUM survey, MeerKAT; arXiv:2603.21845): MIMBH < 105 M☉ at 90% CL. The Häberle lower bound sits well below this ceiling; the two are not in conflict.
A separate, tighter constraint combines stellar kinematics with millisecond pulsar timing accelerations (Bañares-Hernández et al. 2025, A&A 693, A104): a 3σ upper limit of < 6,000 M☉ on any point-mass IMBH, favoring instead an extended dark mass of ~2–3×105 M☉ (a swarm of stellar-mass remnants rather than a single object). This upper limit is directly and formally in tension with the Häberle lower bound of ≥8,200 M☉; the two are not reconcilable under standard modeling assumptions, and this review does not collapse them into a single number.
An earlier N-body analysis (Baumgardt 2017) preferred a much lower mass or no IMBH, placing a 3σ upper limit at ~3,000 M☉. This is in tension with the Häberle detection; the discrepancy is a second, independent unresolved issue in the field and may reflect modeling assumptions in the N-body approach rather than a fundamental conflict.
Current picture: the field carries two unresolved tensions rather than one converged number: Häberle (≥8,200 M☉) vs. Bañares-Hernández (<6,000 M☉), and Häberle vs. the older Baumgardt N-body limit (<3,000 M☉). The IMBH hypothesis is supported if the Häberle kinematic detection is accepted at face value; an extended dark-mass (stellar-mass-BH swarm) model is favored if the Bañares-Hernández combined analysis is weighted instead. LISA gravitational-wave observations are expected to be the eventual arbiter.
All constraints listed are on the mass of the central compact object, unless otherwise noted. CL = confidence level. σ values given are 1σ unless stated.
| Channel | Measurement | CL / σ | Method | Reference | Status |
|---|---|---|---|---|---|
| Stellar kinematics | ≥ 8,200 M☉ | lower bound | 7 fast-moving stars within 0.08 pc; HST + Gaia proper motions | Häberle et al. 2024 (Nature 631:285) | Detection Tension |
| Kinematics + pulsar timing (combined) | < 6,000 M☉ | 3σ upper limit | Joint stellar kinematics + MSP timing acceleration analysis; favors extended dark mass ~2–3×105 M☉ | Bañares-Hernández et al. 2025 (A&A 693, A104) | Upper limit Tension |
| N-body dynamics | < 3,000 M☉ | 3σ upper limit | N-body modeling of velocity dispersion profile | Baumgardt (2017, MNRAS 474) | Upper limit Tension |
| Velocity dispersion | σ0 = 16.8 km/s | — | HST proper motion catalog; mass–σ extrapolation | Baumgardt & Hilker 2018 (MNRAS 478) | Measurement |
| Pulsar timing (alone) | < 105 M☉ | 90% CL | MeerKAT timing of millisecond pulsars; gravitational potential curvature | TRAPUM 2026 (arXiv:2603.21845) | Upper limit |
| IR accretion | Lacc ≪ LEdd | 95% CL | JWST NIRCam/MIRI photometry at the Häberle position; no point source detected | Chen et al. 2025 (ApJ) | Non-detection |
| Radio accretion | ε < 4×10−3 | — | ~170 hr ATCA, 5.5+9.0 GHz; no central point source; bounded accretion efficiency | Mahida et al. 2026 (ApJ 996, 122) | Non-detection |
| X-ray accretion | LX upper limit | — | ~291 ks Chandra ACIS exposure; no central point source detected | Haggard et al. 2013 (ApJL 773:L31) | Non-detection |
| Radio technosignature | EIRP < 1011–1016 W | 95% CL | COSMIC commensal system on the VLA Sky Survey; 950,000+ objects, L+S band; distance 17,900 ly | Sheikh et al. 2025 (AJ, arXiv:2501.17997) | Non-detection |
| Gravitational waves | No CW signal | O3/O4 | LIGO/Virgo O3–O4 continuous GW search; no inspiral signal at OC position | LIGO Collaboration (2023) | Non-detection |
The primary positive evidence for an IMBH comes from Häberle et al. (2024), who identified seven stars within 0.08 pc of the cluster center with proper motions implying velocities up to ~50 km/s, far exceeding what is possible from the stellar mass distribution alone. These instantaneous velocities, exceeding the local escape speed, require a compact dark mass of at least M = 8,200 M☉. This is a velocity-based lower bound, not an orbit fit: orbital curvature (acceleration) of the fast stars has not yet been directly detected, and no closed Keplerian orbit has been measured for any of them. ELT/MICADO astrometry is expected to detect that curvature within a few years of first light.
The critical assumption is that the fast-moving stars are genuinely bound to the cluster center and not unrelated interlopers or binary-ejection remnants. Häberle et al. argue the probability of chance alignment is low given the coherent Keplerian structure in phase space, but this is an unverified claim pending multi-epoch follow-up astrometry (proposed with ELT/MICADO and continued HST/Gaia monitoring).
Epistemic tier: Established observational detection; interpretation requires additional epoch coverage for full orbit verification. The stellar kinematics result is the single most compelling positive evidence for an IMBH in OC.
Baumgardt (2017) ran detailed N-body simulations of OC's velocity dispersion profile and found that the profile is better fit by a model without a central IMBH, or with a mass < 3,000 M☉ at 3σ. This pre-dates the Häberle fast-mover discovery and was based on the global velocity dispersion profile rather than the high-resolution kinematics of the innermost stars.
Baumgardt & Hilker (2018) measured σ0 = 16.8 km/s from HST proper motions, which implies (via the M–σ relation) a central mass consistent with either a small IMBH or a population of stellar remnants. The M–σ relation is calibrated on nucleated galaxies and may not apply to globular clusters.
Key caveat: N-body models are sensitive to the assumed stellar mass function, retention fraction of stellar-mass black holes, and binary fraction near the center. Updated N-body models incorporating the Häberle fast-mover data and modern black-hole retention fractions are needed before the Baumgardt 2017 constraint can be definitively weighed against the kinematic detection.
The TRAPUM survey (TRAnsients And Pulsars Using MeerKAT; Colom i Bernadich et al. 2026, arXiv:2603.21845) placed a 90% CL upper limit of MIMBH < 105 M☉ from the timing of millisecond pulsars in the cluster, using timing alone. This limit comes from the absence of large timing residuals that would indicate a massive perturber near the cluster center.
The Häberle lower bound (≥8,200 M☉) is well below this pulsar-timing-alone ceiling; the two measurements are not in conflict. However, this limit is not yet constraining at the Häberle mass scale on its own. Future observations with the SKA (projected 2030s) and the discovery of additional inner pulsars could push this limit down by one to two orders of magnitude.
A separate analysis, Bañares-Hernández et al. (2025), combines stellar kinematics with millisecond pulsar timing accelerations rather than using timing alone, and reaches a materially different result: a 3σ upper limit of MIMBH < 6,000 M☉, favoring an extended dark mass (~2–3×105 M☉) over a single point mass. This bound is directly below the Häberle lower bound of ≥8,200 M☉ and is not reconcilable with it under standard modeling assumptions; see §4 for discussion. Häberle et al. probe stellar motions within a few arcseconds of the center, while Bañares-Hernández et al. use combined stellar kinematics plus MSP line-of-sight accelerations over a broader radial range; the discrepancy likely reflects these different spatial scales and dynamical assumptions rather than a measurement error in either study.
Epistemic tier: Both are established observational upper limits. The TRAPUM timing-alone limit does not constrain the Häberle detection; the Bañares-Hernández combined limit does, and the tension between them is unresolved.
Chen et al. (2025) searched for an IR point source at the Häberle position using JWST NIRCam and MIRI and found no detection. This places an upper limit on the accretion luminosity Lacc ≪ LEdd — consistent with Bondi accretion in a gas-poor environment, which is expected given OC's depleted interstellar medium. The non-detection does not rule out an IMBH; it merely confirms that any IMBH is accreting well below the Eddington rate, which is the norm for quiescent IMBHs.
Mahida et al. (2026) combined ~170 hr of ATCA observations at 5.5 and 9.0 GHz and detected no central radio point source, bounding the accretion efficiency at ε < 4×10−3. Haggard et al. (2013) obtained a ~291 ks Chandra ACIS exposure and likewise detected no central X-ray point source. Both non-detections are consistent with a quiescent, gas-starved IMBH and do not, by themselves, distinguish the IMBH hypothesis from the extended dark-mass alternative.
The COSMIC system, running commensally on the VLA Sky Survey (Sheikh et al. 2025), searched 950,000+ objects and set EIRP upper limits of 1011–1016 W (depending on frequency and bandwidth assumption) at OC's distance of 17,900 ly. No narrowband technosignature candidates were confirmed. The LIGO/Virgo O3–O4 searches found no continuous gravitational wave signal at the OC position. Both non-detections are consistent with a quiescent IMBH and provide no positive or negative evidence for the IMBH hypothesis.
Future editions will be published quarterly (Q3 2026, Q4 2026, …). Significant new measurements trigger an out-of-cycle patch release (vX.Y). Additions to the measurement ledger require a peer-reviewed or arXiv-posted primary source.
If you use this review in academic work, please cite the Zenodo archive of the OCS toolkit (DOI: 10.5281/zenodo.20689279) and note the review version and date accessed.
BibTeX:
Plain text (APA-style):
Swanson, T. (2026). Living Review: IMBH Evidence in Omega Centauri (NGC 5139) (v1.1, Q3 2026). The Omega Centauri Society. https://doi.org/10.5281/zenodo.20689279
For citing individual interactive tools, use the tool-level BibTeX available on the Cite page.