Cold gas mass measurements for the era of large optical spectroscopic surveys

TL;DR

This paper introduces an indirect method to measure cold gas in galaxies using optical emission lines, calibrated with direct measurements, enabling large-scale studies of gas's role in galaxy evolution.

Contribution

It develops a new calibration for estimating gas surface density from optical data, accounting for scale dependence, and demonstrates its effectiveness in galaxy evolution analysis.

Findings

Indirect gas measurements match direct data after calibration.

Including gas mass reduces scatter in the mass-metallicity relation.

Method enables large sample studies of gas in galaxy evolution.

Abstract

Gas plays an important role in many processes in galaxy formation and evolution, but quantifying the importance of gas has been hindered by the challenge to measure gas masses for large samples of galaxies. Datasets of direct atomic and molecular gas measurements are sufficient to establish simple scaling relations, but often not large enough to quantify three-parameter relations, or second order dependencies. As an alternative approach, we derive here indirect cold gas measurements from optical emission lines using photoionization models for galaxies in the SDSS main galaxy sample and the PHANGS-MUSE survey. We calibrate the gas surface density measurements using xCOLD GASS and PHANGS-ALMA molecular gas measurements to ensure our measurements are reliable. We demonstrate the importance of taking into account the scale-dependence of the relation between optical depth ($\tau_V$) and gas surface density ($\Sigma_{gas}$) and provide a general prescription to estimate $\Sigma_{gas}$ from $\tau_V$, metallicity and the dust-to-metal ratio, at any arbitrary physical resolution. To demonstrate that the indirect cold gas masses are accurate enough to quantify the role of gas in galaxy evolution, we study the mass-metallicity relation (MZR) of SDSS galaxies and show that as a third parameter, gas mass is better than SFR at reducing the scatter of the relation, as predicted by models and simulations.