Low-mass stars dominate the hot (0.7 keV) Galactic X-ray emission

TL;DR

The paper tackles whether the Milky Way's purported super-virial 0.7 keV X-ray emission is primarily from a diffuse hot halo or from coronal emission of ordinary stars. By performing tile-based spectral analyses of the eROSITA soft X-ray background and correlating the 0.7 keV emission with the Milky Way’s stellar mass distribution, the authors show a strong link to low-mass stars, estimating an average X-ray luminosity per stellar mass of $L_{ m x}/M_18=(1.05 ext{--}1.18) imes10^{28}$ erg s$^{-1}$ M$_pped$^{-1}$ and finding that unresolved M dwarfs and FGK stars likely dominate the emission. Bright point sources account for at least ~33% of the flux, and after their removal the stellar contribution remains substantial, with a north–south asymmetry largely attributed to the Sun’s offset from the mid-plane and nearby stellar structures. A test for a diffuse super-virial CGM component via a $eta$-model with $eta=0.4$ yields only an upper limit on the electron density at 10 kpc, $n_e<4 imes10^{-4}$ cm$^{-3}$, suggesting the 0.7 keV emission is not required to originate from a homogeneous halo. Overall, the work provides a robust stellar-origin explanation for most of the 0.7 keV emission and offers a quantitative framework to disentangle stellar versus diffuse CGM contributions in external galaxies.

Abstract

The circumgalactic medium (CGM) of the Milky Way is composed of a tenuous atmosphere filled with multi-phase plasma, including a warm-hot virialised component. Recent studies suggest a much hotter (~0.7 keV) super-virial component detected in both absorption and emission. We want to shed light on the nature of this putative super-virial component. We analysed the X-ray background as observed by SRG/eROSITA over the entire western Galactic hemisphere. We show that low-mass stars provide a large fraction of the 0.7 keV emission. Indeed, a tight correlation is found between the surface brightness of the 0.7 keV emission and the mass distribution of the Milky Way across a large portion of the western Galactic hemisphere. The correlation coefficient implies an X-ray luminosity per unit of stellar mass comparable to that of the average low-mass stars within 10 pc of the Sun, suggesting that unresolved M dwarfs and F, G, and K type stars dominate the 0.7 keV emission. This emission is asymmetric with respect to the Galactic plane, influenced by the asymmetric distribution of nearby star-forming regions, and broadly consistent with the known offset of the Sun above the Galactic midplane. The remaining signal might be produced by the cumulative emission of stars of different types or ages, in addition to other sources (e.g. hot interstellar medium, Galactic corona, etc.). Assuming that the putative residual hot super-virial atmosphere is homogeneous and has a spherical beta profile with slope $β=0.4$, we constrain its density at 10 kpc to be $n_e<4\times10^{-4}$~cm$^{-3}$. Our findings may help refine models of the circumgalactic medium around external galaxies, advancing our understanding of hot baryon flows and galaxy evolution.

Figures (12)

• Figure 1: Left panel: Logarithm of emission measure of the 0.7 keV Galactic emission as observed by eROSITA in the western Galactic hemisphere when fitting the total X-ray emission (diffuse and point sources); it is given in Galactic coordinates and in zenithal equal-area projection centred on ($l=270^\circ$, $b=0^\circ$). The dashed black, red, and blue lines show the longitudinal stripes in Fig. \ref{['fig:Multistripe']}. The dot-dashed white contours indicate the region with the highest concentration of stars within 500 pc of the Sun, as shown in Fig. \ref{['fig:StarFrac']}. Right panel: Same as left panel, but once the stellar contribution, as described by Hunter24, had been removed.
• Figure 2: Best-fit spectra for six selected sky tiles within $220^\circ<l<235^\circ$, within the dashed black lines shown in Fig. \ref{['fig:MapCoro']}. The left column shows spectra within the northern hemisphere at decreasing latitudes. The right column shows spectra at similar latitudes in the southern hemisphere. The black data show the eROSITA spectrum, while the grey points show the flux measured by ROSAT in the softest bands. The solid lines show the contribution from the local hot bubble (red), the warm-hot bubble (blue), the 0.7 keV component (green), the cosmic X-ray background (CXB; magenta), and the instrumental background (black). The inset reports the Galactic latitude and longitude of the centre of the sky tile as well as the total hydrogen column density of neutral absorption derived from the HI4PI survey $N_{\rm H,HI4}$ and the best-fit value, $N_{\rm H,bf}$. The dotted lines show the contribution of the various components of the instrumental background. The lower panels show the fit residuals, with the sky tile number and reduced $\chi^2$.
• Figure 3: Emission measure of 0.7 keV emission for sky tiles within $220^\circ<l<235^\circ$. Top panel: Black and grey data points show the emission measure for the sky tiles with total hydrogen column density of absorption lower and higher than $\log(N_{\rm H,HI4}/{\rm cm^{-2}})>21.5$, respectively. The red and blue curves show the best fit with the mass distribution of the Milky Way as described by Hunter24 and McMillan17, respectively, assuming the Sun to lie on the Galactic plane. The right y-axis in mass surface density is valid for the Hunter24 best-fit normalisation. The vertical dotted grey lines indicate the regions close to the Galactic disc, where the values from the HI4PI survey are in excess of $\log(N_{\rm H,HI4}/{\rm cm^{-2}})>21.5$. The green data show the emission measure of the spectra displayed in Fig. \ref{['fig:spec']}. Bottom panel: Black data show the same emission measures as in the top panel, re-binned over four consecutive (in Galactic latitude) sky tiles. The grey points show the same data, once the Galactic latitudes were flipped in terms of the sign (e.g. inverted northern and southern hemispheres). The solid and dashed red lines show the best-fit relation when the mass distribution is described by Hunter24, with the Sun located on the Galactic plane and 47 pc above it, respectively. The dashed black line shows the best-fit mass distribution as described by McMillan17 for a height of the Sun of 42 pc above the Galactic plane. The right y-axis in mass surface density is valid for the Hunter24 model with $h_\odot=0$ best-fit normalisation.
• Figure 4: Correlation between mass distribution of the Milky Way Hunter24 and total surface brightness, in the 0.1--10 keV band, of the 0.7 keV component within the sky area considered in this work (one data point for each sky tile). The surface brightness was computed extrapolating the best fit 0.7 keV component to the 0.1--10 keV band, and includes all emission, i.e. point sources and diffuse emission. We rebinned the points into N bins along the $x$-axis, evenly spaced in $\log_{10,}$ of the stellar mass surface density. In each bin, the dark blue dot shows the average ($\log_{10}$) surface brightness, while the $x$ and $y$ error bars show the bin's half-width and the standard dispersion of the ($\log_{10}$) surface brightness in the bin, respectively. Each bin is also divided into other M sub-bins along the $y$-axis, with each sub-bin coloured according to the number of encompassed data points, as labelled in the colour bar on the right. The red line shows the relation between surface brightness and stellar mass surface density with $L_x/M = 10.5\times10^{27}$ erg s$^{-1}$ M$^{-1}_\odot$, as derived from the fit (Table \ref{['Tab']}).
• Figure 5: Emission measures of 0.7 keV component and its variations along three longitudinal stripes. The black, red, and blue data points show the emission measure of the 0.7 keV component within the stripes: $220^\circ<l<235^\circ$; $250^\circ<l<265^\circ$; and $295^\circ<l<310^\circ$, respectively, which correspond to the same coloured dotted lines as in Fig. 1. The open grey squares show the emission measure of the 0.7 keV component along the same stripes, once the bright point sources were removed. The coloured and grey lines show the best-fit stellar surface density model (with the Hunter24 model), which reproduces the emission measures where the bright point sources are either retained or removed, respectively. The fit was performed over the entire region considered in this work; it does not consider the addition of a beta model (see Table \ref{['Tab']}). The solid lines show the model's prediction (at the mean longitude of each stripe), which fits the entire region selected for our analysis. Bottom right panel: Best-fit position of the Sun above the Galactic plane ($h_\odot$) as a function of the Galactic longitude derived by fitting the 0.7 keV emission measure along the same three longitudinal stripes. The filled coloured circles and the open grey squares show the best-fit offset of the Sun without and when removing the contribution from the bright point sources. The horizontal shaded blue interval shows the uncertainty on the offset of the Sun above the Galactic plane measured with other tracers Bland-Hawthorn16Griv21.