2003ApJ...587..278W

Much of the interstellar medium in disk galaxies is in the form of neutral atomic hydrogen, H I. This gas can be in thermal equilibrium at relatively low temperatures, T≲300 K (the cold neutral medium [CNM]), or at temperatures somewhat less than 10⁴ K (the warm neutral medium [WNM]). These two phases can coexist over a narrow range of pressures, P_min≤P≤P_max. We determine P_minand P_maxin the plane of the Galaxy as a function of Galactocentric radius R using recent determinations of the gas heating rate and the gas-phase abundances of interstellar gas. We provide an analytic approximation for P_minas a function of metallicity, far-ultraviolet radiation field, and the ionization rate of atomic hydrogen. Our analytic results show that the existence of P_min, or the possibility of a two-phase equilibrium, generally requires that H⁺ exceed C⁺ in abundance at P_min. The abundance of H⁺ is set by EUV/soft X-ray photoionization and by recombination with negatively charged polycyclic aromatic hydrocarbons. In order to assess whether thermal or pressure equilibrium is a realistic assumption, we define a parameter Υ≡t_cool/t_shock, where t_coolis the gas cooling time and t_shockis the characteristic shock time or ``time between shocks in a turbulent medium.'' For Υ<1 gas has time to reach thermal balance between supernova-induced shocks. We find that this condition is satisfied in the Galactic disk, and thus the two-phase description of the interstellar H I is approximately valid even in the presence of interstellar turbulence. Observationally, the mean density n_HI_is often better determined than the local density, and we cast our results in terms of n_HI_as well. Over most of the disk of the Galaxy, the H I must be in two phases: the weight of the H I in the gravitational potential of the Galaxy is large enough to generate thermal pressures exceeding P_min, so that turbulent pressure fluctuations can produce cold gas that is thermally stable; and the mean density of the H I is too low for the gas to be all CNM. Our models predict the presence of CNM gas to R≃16-18 kpc, somewhat farther than previous estimates. An estimate for the typical thermal pressure in the Galactic plane for 3kpc≲R≲18 kpc is P_th/k≃1.4x10⁴exp(-R/5.5kpc) K.cm^–3. At the solar circle, this gives P_th/k≃3000 K.cm^–3. We show that this pressure is consistent with the C I*/C I_totratio observed by Jenkins & Tripp and the CNM temperature found by Heiles & Troland. We also examine the potential impact of turbulent heating on our results and provide parameterized expressions for the heating rate as a function of Galactic radius. Although the uncertainties are large, our models predict that including turbulent heating does not significantly change our results and that thermal pressures remain above P_minto R≃18 kpc.