The Local Group L-band Survey: Probing Cold Atomic Gas in IC10 with Neutral Hydrogen Absorption

TL;DR

This work reports the first localized detections of cold neutral medium (CNM) in IC10 via HI absorption using the Local Group L-band Survey, achieving ~15 pc spatial and 0.4 km s$^{-1}$ spectral resolution. Gaussian decomposition and radiative transfer under a two-phase HI framework yield CNM spin temperatures of roughly $T_s\sim$30–55 K, CNM column densities of $(0.6$–$3)\times 10^{21}$ cm$^{-2}$, and CNM fractions $f_{\rm CNM}\approx 0.21$–$0.37$ along three sightlines, with absorption coinciding in velocity and width with CO, HCO$^+$, and HCN emission, indicating a well-mixed dense atomic and molecular ISM. The detections occur toward internal HII regions near GMCs, showing broad HI lines likely influenced by unresolved CNM substructure and beam blending at ~15 pc scales, and suggesting stronger dynamical coupling between CNM and molecular gas in low-metallicity environments. Comparisons with other Local Group dwarfs and considerations of selection biases imply that IC10 harbors comparatively high per-sightline CNM fractions, though global trends with metallicity require larger, deeper datasets. These results provide empirical constraints on CNM heating/cooling and the atomic-to-molecular gas transition in star-forming, low-metallicity galaxies, informing models of early-Universe ISM conditions.

Abstract

We present the first localized detections of the cold neutral medium (CNM) in IC10, offering a rare view of dense atomic gas in a low-metallicity (0.27 solar metallicity) dwarf galaxy. As a low-metallicity starburst, IC10's interstellar medium conditions could reflect small-scale physical conditions that mirror those of early galaxies, providing a unique window into the heating and cooling processes that shaped the interstellar medium in early-Universe environments. Leveraging the high angular (<5'' ~ 15pc) and spectral (0.4 km/s) resolution of the Local Group L-band Survey, we searched for HI absorption against nine continuum radio sources and detected absorption along three sightlines corresponding to internal radio emission sources within IC10. Using Gaussian decomposition and radiative transfer, we characterize the CNM, deriving spin temperatures of ~30-55 K, column densities of (0.6-3.0)x 1$0^{21}$ cm$^{-2}$, cold HI fractions of ~ 21-37%, and line widths of ~ 5.6-13.6 km/s. For each individual detection of HI absorption, we find corresponding molecular emission from 12CO (J=1-0), HCO+ (J=1-0), and HCN (J=1-0) at similar velocities and with comparable linewidths, indicating a well-mixed cold atomic and molecular medium. In IC10, the CNM shows a clear kinematic connection to the high-density ISM, implying a stronger dynamical coupling with molecular gas than in the Milky Way, in line with expectations for low-metallicity environments. At the ~ 15 pc scales probed by slightly extended HII regions in IC10, unresolved CNM clouds likely contribute to line blending, so the observed broad HI linewidths may partly reflect spatial and kinematic averaging.

Figures (8)

• Figure 1: Peak HI brightness temperature map of IC 10. The 9 radio continuum sources analyzed for HI absorption, a mix of internal and background sources, are marked: 3 internal sources with detected absorption are shown as stars, and 6 non-detections (internal or background) as crosses. For each non-detection, the integrated flux density (in mJy) is labeled. Integrated flux densities for all sources, including the detections, are listed in Table \ref{['tab:observations']}.
• Figure 2: Absorption (top) and emission (bottom) spectra for the three sightlines with detected HI absorption, constructed as described in Section \ref{['sec:absemsextract']}. In both panels, the gray shaded region shows the $1\sigma$ noise level. In the absorption spectra, the red horizontal line indicates the $e^{-\tau}=1.0$ continuum level, and the dashed orange line marks the $3\sigma$ noise threshold. Green shaded regions highlight channels where HI absorption exceeds $3\sigma$.
• Figure 3: Integrated HI intensity map (synthesized beam of $4.22" \times 3.98"$) alongside a wide-band L-band $1.0-2.0$ GHz radio continuum image Corbould2025, with the beams shown in the lower left corner of each panel. H$\alpha$littlethings and CO Corbould2025 contours are overlaid at $3\sigma$, $5\sigma$, and $10\sigma$ levels, shown in purple and teal, respectively. Stars indicate locations of HI absorption detections.
• Figure 4: Peak continuum flux density versus $1\sigma$ noise in $e^{-\tau}$ for absorption detections in IC 10 (pink stars), the SMC from Dempsey (circles, color-coded by peak optical depth $\tau$), and NGC 6822 (green triangles). IC 10 and NGC 6822 absorption spectra were smoothed using a Gaussian kernel to achieve similar spectral resolution to GASKAP. The peak flux density of the LGLBS background sources has been scaled to address the difference in beam widths between the two surveys. IC 10 points lie systematically below the SMC, as the surveyed region contains fewer bright sources. Fluxes and $1\sigma$ noise in $e^{-\tau}$ are scaled to the GASKAP beam and spectral resolution for a fair comparison.
• Figure 5: Emission and absorption spectra for the three sources, along with their Gaussian decompositions and residuals. The top panels display the HI emission spectra, while the bottom panels show the optical depth spectra, plotted as $(1 - e^{-\tau})$. The original spectra are shown in blue, with the best-fit model from radiative transfer calculations (see Section \ref{['fig:radtrans']}) overlaid in orange. In the emission panels, the contribution of each WNM component is plotted in red, and the contribution of each CNM component to the emission is shown in green. In the absorption panels, the contribution of each individual CNM component is shown in green.The fitting parameters are provided in Table \ref{['tab:radtrans']}