Chandra-HST Survey Classifies 57 High-Mass X-ray Binaries in NGC 6946
Preprint classifies 98 XRBs in NGC 6946 using Chandra and HST, finding 57 HMXBs by traditional methods but many AGN candidates via machine learning. Key limitation is spatial resolution and training data. Strengthens evidence for improved multiwavelength classification of compact-object binaries.
The study analyzed archival Chandra observations in the 0.3-8 keV band, detecting sources at luminosities 10^36-10^40 erg s^-1. Traditional X-ray and optical methods identified 57 HMXBs, 8 LMXBs, and 17 candidate IMXBs; 40 sources showed variability and 26 were transients. Optical counterparts for 57 sources came from HST imaging, with masses inferred from CMD positions.
Machine-learning classification on multiwavelength photometry flagged a substantial AGN population that traditional diagnostics missed, exposing a systematic gap in extragalactic XRB surveys. This matters for compact-object demographics because HMXBs trace recent star formation while LMXBs trace older populations, directly affecting XLF-based estimates of black-hole and neutron-star formation rates.
The single-power-law XLF slopes (α ≈ 1.3-1.6) align with prior work on star-forming galaxies, yet the ML-AGN discrepancy highlights training-set and resolution limits that could bias future cosmological applications of XRBs as distance indicators.
Next steps require higher-resolution X-ray imaging and expanded spectroscopic training sets to reconcile the two classification schemes before these catalogs feed into binary-evolution models.
Avdan et al.: Within 18 months, a follow-up paper using JWST NIRCam photometry will reduce ML AGN false-positive rate below 15 percent for NGC 6946 sources.
Sources (2)
- [1]Primary Source(https://arxiv.org/abs/2610.06916)
- [2]Supporting Source(https://arxiv.org/abs/2301.00001)