2024ApJ...961..161M

We develop a new method for spatially mapping a lower limit on the mass fraction of the cold neutral medium by analyzing the amplitude structure of ${\hat{T}}_{b}({k}_{v})$, the Fourier transform of T_b(v), the spectrum of the brightness temperature of the H I 21 cm line emission with respect to the radial velocity v. This advances a broader effort exploiting 21 cm emission line data alone (without absorption line data, τ) to extract integrated properties of the multiphase structure of the H I gas and to map each phase separately. Using toy models, we illustrate the origin of interference patterns seen in ${\hat{T}}_{b}({k}_{v})$. Building on this, a lower limit on the cold gas mass fraction is obtained from the amplitude of ${\hat{T}}_{b}$ at high k_v. Tested on a numerical simulation of thermally bi-stable turbulence, the lower limit from this method has a strong linear correlation with the “true” cold gas mass fraction from the simulation for a relatively low cold gas mass fraction. At a higher mass fraction, our lower limit is lower than the “true” value, because of a combination of interference and opacity effects. Comparison with absorption surveys shows a similar behavior, with a departure from linear correlation at N_H I_ ≳ 3-5 × 10²⁰ cm^–2. Application to the DRAO Deep Field from DHIGLS reveals a complex network of cold filaments in the Spider, an important structural property of the thermal condensation of the H I gas. Application to the HI4PI survey in the velocity range -90 < v < 90 km s^–1 produces a full sky map of a lower limit on the mass fraction of the cold neutral medium at 16.'2 resolution. Our new method has the ability to extract a lower limit on the cold gas mass fraction for massive amounts of emission line data alone with low computing time and memory, pointing the way to new approaches suitable for the new generation of radio interferometers.