Due to their high stellar density and large population of stellar remnants, globular clusters (GCs) are a unique environment for the efficient formation of low mass X-ray binaries (LMXBs), systems with a neutron star (NS) or stellar-mass black hole (BH) accreting matter from a main sequence or red giant star through Roche-lobe overflow. The pathways to LMXB formation in GCs include tidal capture, 3-body binary exchange, and direct stellar collisions with compact objects, while field LMXBs likely evolved as isolated binaries. The formation of LMXBs through close encounters, possibly through diverse channels, accounts for the high specific abundance of LMXBs in GCs compared to the field both in the Milky Way and in other galaxies (e.g., Pooley et al. 2003; Kundu et al. 2002).

While a substantial fraction of field LMXBs in the Milky Way host BHs (e.g., Tetarenko et al. 2016), the overwhelming majority of LMXBs in GCs—at least those for which classification of the compact object has been possible—host NSs rather than BHs (Bahramian et al. 2014). This was originally explained with analytic arguments about the fate of a population of BHs in the dense cluster environment. After formation, it is expected that cluster BHs will sink to center of the cluster and become mass segregated from the other, less massive cluster members. In these close quarters, the BHs should form tight binaries that are largely ejected through interactions with other BHs or BH–BH binaries. This process was argued to continue until all, or nearly all BHs were depleted from the GC (Sigurdsson & Hernquist 1993; Kulkarni et al. 1993).

Parallel observational and theoretical tracks have led to a reconsideration of this conclusion. Several GCs in external galaxies may contain BHs accreting near the Eddington luminosity, with the quality of the evidence ranging from suggestive to compelling (e.g., Maccarone et al. 2007; Zepf et al. 2008; Irwin et al. 2010; Peacock et al. 2012). In the Milky Way, low-luminosity BH candidates have been identified by a combination of radio continuum, X-ray, and optical data in the GCs M22 (Strader et al. 2012), M62 (Chomiuk et al. 2013), and 47 Tuc (Miller-Jones et al. 2015; Bahramian et al. 2017). A number of theoretical papers have concluded that BH ejection is less efficient than originally thought, since a putative subcluster of BHs cannot remain dynamically isolated from the rest of the GC as its mass declines (Sippel & Hurley 2013; Morscher et al. 2013; Breen & Heggie 2013; Heggie & Giersz 2014; Morscher et al. 2015). This work has accelerated since the discovery of merging BH–BH binaries by Advanced LIGO (Abbott et al. 2016), as the dynamical formation of BH–BH binaries in GCs may be an important or even dominant channel for such systems (Rodriguez et al. 2016; Chatterjee et al. 2017)

As noted above, no transient LMXB in a Galactic GC has ever been identified to host a BH. Given the small number of transient LMXBs in GCs, this could be a statistical fluctuation, or could reflect different orbital properties for dynamically formed systems: for example, short-period BH LMXBs could undergo shorter, less luminous outbursts that would not be detected by all-sky X-ray monitors (Knevitt et al. 2014). In any case, BH LMXBs are expected to spend most of their evolution in a low-luminosity quiescent state with an X-ray luminosity in the range \(L_{X}\sim 10^{30}\)\(10^{33}\) erg s\({}^{-1}\). In this state it is typically not possible to separate them from other X-ray sources, such as NS LMXBs, accreting white dwarfs, or even active binaries, on the basis of X-ray observations alone. However, in quiescence, BHs are observed to emit steady flat-spectrum radio continuum emission which is thought to originate via partially self-absorbed synchrotron radiation from compact jets (Blandford & Königl 1979). Normal NS LMXBs may also have outflows or jets at these X-ray luminosities, but have not been detected in the radio, and are at least one to two orders of magnitude fainter in radio luminosity at a fixed \(L_{X}\) than BHs (Tudor et al. 2017). A variety of other sources may sometimes show radio continuum emission, which we discuss below in §XXX.

The possibility of identifying quiescent BH LMXHs through radio continuum emission motivated our group to conduct a systematic survey of 50 Galactic GCs using radio continuum observations from the upgraded Karl G. Jansky Very Large Array (VLA) and the Australia Telescope Compact Array (ATCA). We name this survey MAVERIC (Milky-way ATCA and VLA Exploration of Radio-sources in Clusters).

Here we present a multi-wavelength study of a radio-selected BH candidate in the Milky Way GC M10 (NGC 6254; D = 4.4 kpc; Harris 1996). In Section 2, we discuss our VLA observations, Chandra X-ray data, Hubble Space Telescope optical photometry, and ground-based SOAR spectroscopy of the system. In Section 3 we discuss the properties of the binary: identity of the binary companion, orbital parameters, and mass constraints. In Section 4 we discuss the interpretation of the system, and summarize our findings in Section 5.