A Chandra Survey of Milky Way Globular Clusters. I. Emissivity and Abundance of Weak X-Ray Sources

Abstract

Based on archival Chandra data, we have carried out an X-ray survey of 69, or nearly half the known population of, Milky Way globular clusters (GCs), focusing on weak X-ray sources, mainly cataclysmic variables (CVs) and coronally active binaries (ABs). Using the cumulative X-ray luminosity per unit stellar mass (i.e., X-ray emissivity) as a proxy of the source abundance, we demonstrate a paucity (lower by 41% ± 27% on average) of weak X-ray sources in most GCs relative to the field, which is represented by the Solar neighborhood and Local Group dwarf elliptical galaxies. We also revisit the mutual correlations among the cumulative X-ray luminosity (L X ), cluster mass (M ) and stellar encounter rate (Γ), finding L X ∝ M 0.74±0.13 , L X ∝ Γ 0.67±0.07 and Γ ∝ M 1.28±0.17 . The three quantities can further be expressed as L X ∝ M 0.64±0.12 Γ 0.19±0.07 , which indicates that the dynamical formation of CVs and ABs through stellar encounters in GCs is less dominant than previously suggested, and that the primordial formation channel has a substantial contribution. Taking these aspects together, we suggest that a large fraction of primordial, soft binaries have been disrupted in binary-single or binary-binary stellar interactions before they can otherwise evolve into X-ray-emitting close binaries, whereas the same interactions also have led to the formation of new close binaries. No significant correlations between L X /L K and cluster properties, including dynamical age, metallicity and structural parameters, are found.

The X-ray emissivity of low-density stellar populations

“…projected onto low-density populations, we feel it is crucial to undertake careful investigation of each source. In section 5 below, we discuss how a different curation of data have led Cheng et al (2018) to significantly different conclusions than ours. In this section, we investigate three intriguing claims in the literature, about X-ray sources in (the direction of) the low-density populations of the open cluster NGC 6819, the sparse globular cluster NGC 6366, and the Sculptor dwarf galaxy.…”

“…ular clusters. Cheng et al (2018) also find significantly different results; they find a poorer correlation between X-ray emissivity (measured as L X /L K ) and the stellar interaction rate per unit mass (e.g. their fig.…”

“…We suggest some systematic errors that may affect their estimates of cluster mass (using Kband estimates) and of L X . Cheng et al (2018) estimate cluster masses using Kband estimates, while we use mass estimates from Baumgardt & Hilker (2018). We compare the spread in the ratio of Cheng et al's K-band infrared luminosity estimates to mass, with the spread in the ratio of visible-light luminosity (Harris 1996(Harris , 2010 to mass, in Fig.…”

“…We approach this problem by calculating X-ray emission from each cluster by both methods-identifying emission from individual sources (and measuring their membership), and using the X-ray emission enclosed within the entire half-mass radius. In most cases, we use estimates for X-ray emission within the half-mass radius from Cheng et al (2018), and convert them to 0.5-2 keV using either the power-law photon index best fit of Cheng et al (2018), or a photon index of 2. For clusters not studied by Cheng et al (2018), we perform this analysis ourselves.…”

“…In most cases, we use estimates for X-ray emission within the half-mass radius from Cheng et al (2018), and convert them to 0.5-2 keV using either the power-law photon index best fit of Cheng et al (2018), or a photon index of 2. For clusters not studied by Cheng et al (2018), we perform this analysis ourselves. When the two estimates (from adding up point sources, or extracting the full flux within the half mass radius) disagree, we encompass both estimates within our error bars.…”