HEATING OF H-II REGIONS WITH APPLICATION TO THE GALACTIC-CENTER

Recent observations of the H76alpha (Schwarz, Bregman, & van Gorkom) and H92alpha (Roberts et al.) radio recombination lines from the northern arm and bar of Sagittarius A using the Very Large Array (VLA) have suggested that there is a sharp drop in the line-to-continuum ratio in the bar close to the compact nonthermal source Sgr A*. If LTE is assumed, this implies that there is a dramatic increase of the electron temperature, with temperatures rising to T(e) approximately 50,000 K at r approximately 0.25 pc from Sgr A*. In this paper we review the heating and thermal equilibrium of photoionized gas and show that such high temperatures are implausible. Photon-heating mechanisms (UV photoionization heating, grain photoelectric heating, and X-ray heating) either fail to provide the required heating rates or else require that the ionization state of the gas is very high. Gas which is collisionally ionized and heated in shocks can reach high temperatures, but only over a limited range of column densities. Specific application to the Galactic center observations show that the total heating power required to maintain the gas at the derived temperatures, using the observed emission measure in the bar and the temperature distribution derived from the radio recombination lines, is approximately 7 x 10(6) L., comparable to the bolometric luminosity of the central source as measured by the far-infrared flux from grains. Thus, the cooling emission from this hot gas, if the LTE-derived temperatures are correct, would supply a major fraction of the bolometric and ionizing luminosity inferred from the ionized gas in the central 1 pc cavity and the dust and neutral gas in the surrounding torus. Photoionization and grain photoelectric heating fail as sources of the heating power by large factors. While X-ray heating can produce sufficiently high temperatures, it also violates the constraints on the ionization level of the gas in the bar and requires X-ray fluxes exceeding the observed upper limit for Sgr A* by orders of magnitude. Shock-heating can easily produce high gas temperatures, but the column density of gas at temperatures of a few times 10(4) K in shocks is small, requiring a very ad hoc superposition of a large number of shocks, which must also be truncated in column density so the emission measure is not dominated by the 10(4) K postshock gas. A number of more speculative heating mechanisms have also been considered, but are either inadequate or so ad hoc as to be implausible. Failure to find any reasonable mechanism for maintaining the gas at these temperatures, plus the disagreement between the observed Bralpha-to-radio-continuum ratio and that predicted for T(e) approximately 50,000 K gas, provide strong evidence that the gas is in fact at T(e) almost-equal-to 10,000 K and that either observational error or another physical mechanism has produced the low line-to-continuum ratios. We find no physical mechanism that is not ruled out by observations. Nevertheless, we feel that the evidence favors a normal electron temperature (T(e) approximately 10(4) K) for the bar. We propose several additional probes of this mysterious region.