2023MNRAS.521.4323E

We report spectroscopic and photometric follow-up of a dormant black hole (BH) candidate from Gaia DR3. The system, which we call Gaia BH2, contains a ∼1 M_⊙ red giant and a dark companion with mass $M_2 = 8.9\pm 0.3\, {\rm M}_{\odot }$ that is very likely a BH. The orbital period, P_orb = 1277 d, is much longer than that of any previously studied BH binary. Our radial velocity (RV) follow-up over a 7-month period spans >90 per cent of the orbit's RV range and is in excellent agreement with the Gaia solution. UV imaging and high-resolution optical spectra rule out plausible luminous companions that could explain the orbit. The star is a bright (G = 12.3), slightly metal-poor ($\rm [Fe/H]=-0.22$) low-luminosity giant ($T_{\rm eff}=4600\, \rm K$; $R = 7.8\, R_{\odot }$; $\log \left[g/\left({\rm cm\, s^{-2}}\right)\right] = 2.6$). The binary's orbit is moderately eccentric (e = 0.52). The giant is enhanced in α-elements, with $\rm [\alpha /Fe] = +0.26$, but the system's Galactocentric orbit is typical of the thin disc. We obtained X-ray and radio non-detections of the source near periastron, which support BH accretion models in which the net accretion rate at the horizon is much lower than the Bondi-Hoyle-Lyttleton rate. At a distance of 1.16 kpc, Gaia BH2 is the second-nearest known BH, after Gaia BH1. Its orbit - like that of Gaia BH1 - seems too wide to have formed through common envelope evolution. Gaia BH1 and BH2 have orbital periods at opposite edges of the Gaia DR3 sensitivity curve, perhaps hinting at a bimodal intrinsic period distribution for wide BH binaries. Dormant BH binaries like Gaia BH1 and Gaia BH2 significantly outnumber their close, X-ray bright cousins, but their formation pathways remain uncertain.