The Optical Counterparts of the Luminous X‐Ray Binary Stars in Globular Clusters

Abstract

Ten percent of our GalaxyIs luminous ergs (L X 1036 s~1) X-ray point sources are located in globular clusters (GCs), but GCs contribute a far smaller fraction of normal stars to the Galaxy. X-ray bursts have been observed from nearly all GC sources, indicating that they are low-mass X-ray binary (LMXB) systems containing a neutron star and a low-mass companion star from which matter is being transferred. The understanding of LMXB overabundance in GCs may well lead to important insights into the formation and evolution of these exotic binary systems as well as the dynamics of GCs themselves. The goals of this dissertation are to identify the optical counterparts to some GC LMXBs lacking candidates, bring together and compare in the most homogeneous fashion all available Hubble Space T elescope (HST ) optical observations of the current group of GC LMXB counterparts, and discuss the implications for cluster dynamics and LMXB systems in general. In this work, candidates for three additional optical counterparts to luminous X-ray binaries in GCs are presented, thereby doubling the number of optical counterpart candidates. Two, those in NGC 1851 and NGC 6441, are very likely correct although they require additional work to conÐrm the identiÐcations, while the third, in NGC 6652, remains somewhat tentative because of the positional discrepancy with the X-ray coordinates and the fact that the entire error circle is not surveyed. A homogeneous set of HST photometric measurements for all the counterparts identiÐed thus far is presented, and their optical properties are compared with those of Ðeld LMXBs. The mean color of the cluster sources is (U[B)0 identical to that of the Ðeld sources, and the mean is M B0 similar in range, but fainter on average than bursters in the Ðeld. The ratio of optical to X-ray Nux of cluster sources seems to show a signiÐcantly larger dispersion than that of the Ðeld sources. The GC LMXBs seem to conform well to the previously suggested optical luminosity, X-ray luminosity, orbital period relation (J. van Paradijs & J. McClintock, A&A, 290, 133 [1994]). If the source in NGC 1851Efor which no orbital period has yet been determinedEfollows this pattern, then the period will likely be less than 0.8 hr. Similarly, if the object advanced here in NGC 6652 is indeed the correct optical counterpart, this relation suggests that its orbital period will also prove to be less than 1 hr. New and archival HST spectra and imaging data are analyzed to intercompare the UV/optical spectral energy distributions (SEDs) of GC LMXBs. A set of simple model SEDs is introduced and compared with the observations to infer accretion rates, disk diameters, and other properties of these systems. The GC LMXBs are particularly well suited for SED modeling as the distance and reddening of these systems is quite well understood, unlike most Ðeld LMXBs. The sources in NGC 6441 and M15 are found to be optically quite luminous, with comparatively shallow SEDs. The faint sources in NGC 6712 and NGC 1851 have signiÐcantly steeper SEDs and are most likely underluminous because their orbital separations and thus accretion disks are small. The case of NGC 6624 appears slightly di†erent from the others, with an average optical luminosity, yet with the steepest SED slope. This is most likely due to its high accretion rate isD10 times higher than any of the other (L X GC LMXBs) and extremely small size. From its luminosity, it is inferred that the source in NGC 6652 will have similar system parameters and SED to the one in NGC 1851. If these inferences prove correct, it will mean that twothirds of GC LMXBs for which orbital periods are known have periods less than 1 hr and are therefore quite likely double degenerate ultracompact binaries. Such a high fraction of ultracompact systems would be remarkable and may indicate that the processes responsible for creating such short-period systems are more common than previously thought.