Kinematics of H I and O VI Absorbers: Insights into the Turbulence Driver of the Multiphase Circumgalactic Medium

TL;DR

This study uses a large, low-redshift CGM absorber compilation to connect line widths and column densities of H I and O VI to the underlying turbulence driving the multiphase CGM. By modeling the N–b relations under two limiting conditions—constant total hydrogen column $N_{\rm H}$ for H I and roughly constant density $n_{\rm H}$ for O VI—the authors reveal a consistent picture: H I traces a broad range of densities at fixed $N_{\rm H}$, while O VI traces gas at nearly fixed density, with non-thermal motions following a Kolmogorov-like cascade across scales from ~1 pc to ~100 kpc. The O VI multi-component analysis further implies a line-of-sight filling factor of ~25%, suggesting a substantial, but incomplete, volume filling of turbulent gas in halos. Collectively, the results imply that turbulence—driven by halo accretion at large radii—couples CGM phases and sustains the multiphase structure across five orders of magnitude in scale, providing a unified framework for CGM kinematics across environments. This work highlights the importance of combining absorption-line kinematics with extensive galaxy samples to probe the energetics and scale-dependent dynamics of the CGM.

Abstract

We investigate large-scale gas kinematics in the multiphase circumgalactic medium (CGM) using the observed correlation between line width and column density for H I and O VI absorbers. Leveraging extensive public galaxy survey data at $z\lesssim0.1$, we construct a new galaxy sample based on the availability of background QSOs with far-ultraviolet spectra from the Far Ultraviolet Spectroscopic Explorer (FUSE). By combining this FUSE-galaxy sample with literature collections, we find that H I absorbers exhibit a clear inverse correlation between Doppler width and column density over nearly five orders of magnitude in $N_{\rm HI}$, from $N_{\rm HI} \approx 10^{13}\rm~{cm^{-2}}$ to $N_{\rm HI} \approx 10^{18}\rm~{cm^{-2}}$, while O VI absorption follows a positive correlation across $N_{\rm OVI}\approx 3\times10^{13}$-$10^{15}\rm~{cm^{-2}}$. We develop a model framework to interpret these contrasting trends and show that H I absorbers are best described as systems of approximately constant total column density ($N_{\rm H}$), whereas O VI traces regions of roughly constant spatial density ($n_{\rm H}$ and $n_{\rm OVI}$). Under the latter scenario, the observed $b_{\rm OVI}$-$N_{\rm OVI}$ relation maps directly to a velocity-size relation consistent with a Kolmogorov-like turbulent spectrum. Together, these findings reveal a coherent physical picture in which H I and O VI trace a continuous turbulent cascade spanning more than five orders of magnitude in spatial scale-from cool, photoionized clumps to warm, highly ionized halo gas--with accretion in the halo outskirts likely driving the turbulent energy injection that sustains the multiphase CGM.

Figures (5)

• Figure 1: The redshift and projected distance between galaxy and QSO for the FUSE sample.
• Figure 2: The $N$-$b$ relation for individual H1 components, showing a 8.0 $\sigma$ negative correlation. The adopted samples include the FUSE sample (this work), COS-Halos Werk2013, COS-LRG Chen2018Zahedy2019, and CUBS compiled in Qu2022, which are summarized in Section \ref{['sec:data']}. A kinematical model is presented as dashed (constant-$N_{\rm H}$) and dotted lines (constant-$n_{\rm H}$; see details in Section \ref{['sec:h1']}). Except for those systems with $\hbox{${N}_{\rm HI}$}\gtrsim 10^{18}~\hbox{${\rm cm^{-2}}$}$, the majority of the CGM absorbers can be explained by our kinematic model in the range of $\log \hbox{${n}_{\rm H}$}/\hbox{${\rm cm^{-3}}$} \approx -2$ to $-5$ and $\log \hbox{${N}_{\rm H}$}/\hbox{${\rm cm^{-2}}$} \approx 17$-19.
• Figure 3: The $N$-$b$ relation for individual O6 components, showing a 5.1 $\sigma$ positive correlation. The adopted samples are from the FUSE sample (this work), COS-Halos Werk2013, COS-LRG Zahedy2019, and CUBS CUBSVII. Panel (A): A power law fitting suggests a slope of $0.32\pm0.06$ and an intrinsic scatter of $0.16\pm 0.01$ dex. The dotted red lines show the 1 $\sigma_{\rm p}$ and 2 $\sigma_{\rm p}$ scatters associated with the best-fit model as the solid red line, while the dashed black line represents a power law model with a slope of $1/3$. Panel (B): The model-predicted $N$-$b$ relation assuming different $n_{\rm OVI}$ and two fiducial temperatures, $\log T/{\rm K} = 4.5$ under the assumption of PIE (dashed curves) and $\log T/{\rm K} = 5.5$ under the assumption of O6 originating in collisionally-ionized gas (solid curves). In addition, a Kolmogorov-like size-linewidth relation is adopted for modeling the non-thermal contribution to the observed $b_{\rm OVI}$ following a power-law slope of $1/3$ from Chen2023. Panel (C): Expectations from the Stern2018 model for O6, illustrating the correlation between $N$ and $\hbox{${b}_{\rm NT}$}$. As in Panel (B), the solid curves represent collisionally ionized O6 with $\log T/{\rm K} = 5.5$, while the dashed curves correspond to photoionized gas with $\log T/{\rm K} = 4.5$. In both panels (B) and (C), the flattening of the model curves toward low $b$ values reflects the diminishing contribution from non-thermal line width and the increasing dominance of thermal line widths determined by the gas temperature.
• Figure 4: The observed line-of-sight O6 $N_{\rm tot}$-$\sigma_\varv$ relation integrated across all components. Circles represent the results from absorbers showing multiple O6 components, while squares represent systems with a single detected component. Panel (A): A power-law fit to the multi-component sightlines (red circles) yields a best-fit slope of $\alpha=0.31_{-0.12}^{+0.13}$ (see Equation \ref{['eqn:linear']} for the definition of $\alpha$). The solid line and shaded region represent the best-fit model and its associated 1 $\sigma$ uncertainties, while the dotted lines indicate the 1 $\sigma_{\rm p}$ intrinsic scatter. The best-fit model is shown as a red dashed line for $N_{\rm tot, OVI}<10^{14}\,\hbox{${\rm cm^{-2}}$}$ to indicate an extrapolation below the column density range covered by empirical observations for the multi-component systems. The horizontal dotted line marks the expected thermal broadening at $\log T/{\rm K} = 5.5$. The best-fit model for single O6 components from Figure \ref{['fig:o6_Nb']}(A) is duplicated here as the black dashed line. At a given $N_{\rm tot}$, the observed $\sigma_\varv$ values are $\approx 0.2$ dex higher in multi-component absorbers than in single-component systems, implying a filling factor of 25$\%$ along individual sightlines (see details in Section \ref{['sec:o6s']}). For comparison, we also include the expectations from the cooling-flow models in Panel (B) with the solid and dashed-dotted lines showing model expectations for gas of $\log T/{\rm K} = 5.5$ and 6, respectively Heckman2002Bordoloi2017.
• Figure 5: The observed relation between line width and spatial scale in the CGM, combining O6 measurements from this study (red and orange shaded bands) with previous results based on individual absorbing components from Chen2023 (cyan squares). The width and height of the red and orange bands represent the ranges in spatial scale and intrinsic scatter ($\pm\,2\,\sigma_{\rm p}$) for individual and integrated O6 absorbers, respectively (see discussion in §§ \ref{['sec:o6']}–\ref{['sec:o6s']}). The Kolmogorov-like turbulent model derived in Chen2023 is shown as a thin solid line for $l\!<\!1$ kpc and as a dashed line beyond this range, with the associated $1\,\sigma$ scatter indicated by dotted lines. For comparison, the thick solid line shows the expected correlation between projected virial velocity ($\varv_{\rm vir}$) and halo radius ($\hbox{$r_{\rm vir}$}$) for gravitationally bound dark-matter halos with $\log M_{\rm halo}/\hbox{${\rm M}_\odot$}\!=\!10$--13. The relation is dotted below this mass range, where galaxies become too faint to be systematically detected in current surveys Chen2020. The observed line widths on scales of $\sim\!100$ kpc are consistent with gravitational motions within halos, whereas on smaller scales the non-thermal motions follow a Kolmogorov-like turbulent cascade.