High$-z$ [OI] emission lines: ColdSIM simulations and ALMA observations

TL;DR

This work predicts high-redshift [OI] 63 μm and 145 μm emission by post-processing ColdSIM cosmological simulations with the DESPOTIC code, explicitly modeling foreground self-absorption in [OI] 63 μm and employing Random Forests to accelerate predictions. The results show strong correlations between [OI] luminosities and both star formation and molecular gas content, with [OI] 63 μm self-absorption reducing observed flux by a factor of about 2–4 and the [OI] 145 μm line remaining largely optically thin. The predicted line ratios, such as [OI] 63 μm / [CII] 158 μm ≈ 1 and [OI] 145 μm / [CII] 158 μm ≈ 0.2, closely match z ≳ 6 observations when self-absorption is included, supporting the use of [OI] lines as tracers of the early ISM and molecular reservoirs. The study highlights the diagnostic power of joint [OI] detections for constraining self-absorption and ISM conditions during the epoch of reionization, and provides a framework for applying these predictions to upcoming ALMA observations.

Abstract

Neutral-oxygen [OI] far-infrared emission lines at $63μ$m and $145μ$m are powerful probes of the physical conditions in the interstellar medium, although they have not been fully exploited in high-redshift studies. We investigate the connection between [OI] emission lines and key galaxy properties, such as star formation rate (SFR) and H$_2$ content. Our predictions are compared with existing observations and new data analysed in this work. We post-process the outputs of the ColdSIM cosmological simulations with the DESPOTIC model, taking into account [OI]$63μ$m self-absorption by cold foreground material. A Random Forest algorithm is employed to accelerate computations and new observational ALMA data for galaxies at redshift $z\simeq 5-7$ are used to validate our model. Our predictions show significant [OI]$63μ$m luminosities ($\approx 10^8\,\rm L_\odot$) for galaxies with SFRs of $\approx 10^2\,\rm M_\odot\,{\rm yr}^{-1}$. The $145μ$m line luminosity is typically $15 \%$ the [OI]$63μ$m one and is a factor $\approx 2-20$ below high-$z$ observations. Both [OI] lines correlate with SFR and molecular mass, but exhibit flattening in scaling relations with metallicity and stellar mass. Foreground self-absorption reduces the [OI] flux by a factor of $2-4$, consistent with empirical corrections in observational studies. We find typical line ratios of [OI]$63μ$m / [CII]$158μ$m $\approx 1$ and [OI]$145μ$m / [CII]$158μ$m $ \approx 0.2 \, -$ consistent with $z\gtrsim 6$ observations, but only when [OI]$63μ$m self-absorption is included. Both [OI]$63μ$m and [OI]$145μ$m lines serve as tracers of star formation and molecular gas at high redshift. Their joint detection can provide constraints on the properties of the early interstellar medium and self-absorption of the [OI]$63μ$m line.

Figures (14)

• Figure 1: [OI] 145$\mu$m emission maps. Contour lines correspond to [2, 3, 4, 5, 6, 8, 10, 15, 20, 30]$\,\sigma$ confidence levels, where $\sigma = 0.030$, 0.098, 0.082 and 0.024 Jy beam$^{-1}$ km s$^{-1}$ for panels a, b, c, and d, respectively. In panel c, arrows indicate the location in which we detect [OI] or [CII] emission at $z=4.680$, in addition to continuum emission (Figure \ref{['fig:hlsmaps']}). In each panel, the bottom-left white ellipse shows the ALMA beam.
• Figure 2: Visualization of the ${\rm [OI]}\, 63 {\mu \rm{m}}$ and ${\rm [OI]}\, 145 {\mu \rm{m}}$ luminosities, and the ${\rm [OI]}\, 63 {\mu \rm{m}}/{\rm [CII]}\, 158 {\mu \rm{m}}$ and ${\rm [OI]}\, 145 {\mu \rm{m}}/{\rm [CII]}\, 158 {\mu \rm{m}}$ ratios, for the most massive simulated galaxy at $z=6.14$. Black contours follow the gas density distribution.
• Figure 3: Fraction of self-absorbed $63 \, {\mu \rm{m}}$ [OI] line flux due to foreground material as a function of the galaxy SFR as predicted by our model. Solid lines refer to median trends while the shaded area are the 16-84th percentiles dispersion. The solid line refers to the fiducial model, while the dashed one is obtained without considering the condition on velocity in Eq. \ref{['eq:velcond']}. The yellow dashed region represent the typical values adopted in observational studies (reduction by a factor of $\sim 2-4$).
• Figure 4: Luminosity function of ${\rm [OI]}\, 63 {\mu \rm{m}}$ (top panel) and ${\rm [OI]}\, 145 {\mu \rm{m}}$ (bottom panel) at $z=6,\,8,\,{\rm and} \, 10$ as predicted by our simulation. Errorbars are obtained by bootstrapping. The colored arrows in the top panel indicate the mean magnitude of the foreground self-absorption at each redshift. Low opacity points are those affected by resolution effects.
• Figure 5: Relations between the ${\rm [OI]}\, 63 {\mu \rm{m}}$ (red) and ${\rm [OI]}\, 145 {\mu \rm{m}}$ (blue) line luminosities and stellar mass (left), gas-phase metallicity traced by oxygen-to-hydrogen number ratio with respect to solar (middle), and H$_2$ mass (right) for our simulated galaxies at $z = 6$. Solid lines represent the median trends, while density contours illustrate the underlying distribution of galaxies.