HII regions rarely display the spherical symmetry assumed by the standard Strömgren theory of an expanding volume of photoionized gas. Rather, density gradients between molecular clouds, where massive stars are born, and their surroundings cause large distortions in the shapes of HII regions, and flows of material towards the less dense regions (see for example Gómez et al. 1993, Wilcots 1994, Keto et al. 1995, Russeil et al. 1995, Koo et al. 1996). Such density gradients have been invoked to explain weakly collimated outflows produced by the winds of low mass stars, and were numerically modelled by Róyczka & Tenorio-Tagle 1985a, 1985b. The evolution of HII regions in the presence of density discontinuities was studied by Tenorio-Tagle and coworkers in a series of papers (Tenorio-Tagle 1979, 1982, Tenorio-Tagle & Bedijn 1982, Tenorio-Tagle et al. 1979, Bodenheimer el al. 1979) where the bases of the so-called champagne model was established. Some features of this model, such as the runaway expansion of an ionization front, are also reproduced in outward decaying, spherically symmetric density distributions, provided that the density gradient exceeds some critical value (Franco et al. 1990, Rodríguez- Gaspar et al. 1995). The champagne model achieved a remarkable success in explaining the general morphology of actual HII regions and features in their velocity fields (see review by Yorke 1986).
The champagne model considers photoionization by a central star as the source of energy behind the expansion. Nevertheless, it is widely recognized now that the strong stellar winds produced by the same massive stars which ionize the HII regions play a major role in the large scale shaping of the interstellar medium by generating expanding structures around them (Weaver et al. 1977, Bruhweiler et al 1980, Beltrametti et al. 1982, Tenorio-Tagle et al. 1982, Tenorio-Tagle & Bodenheimer 1988, Mac Low et al. 1989, Bisnovatyi-Kogan & Silich 1995). Although most of the energetic output of young massive stars is in the form of ultraviolet radiation, it turns out that the mechanical power contained in the stellar wind is more efficiently transferred to the surrounding gas (Dyson & Williams 1980), making the energetic output from both sources of similar importance. Visible-light photographs of HII regions often display bubble-like structures in the lower density areas reminiscent of the effects of the low density, hot gas produced by shocked stellar winds, rather than the denser gas flows expected from the champagne model (see e.g. photograph of NGC 6357 in Malin 1993, p. 91). Stellar winds also play a fundamental role in explaining the basic features of ultracompact HII regions (Van Buren et al. 1990; Mac Low et al. 1991; Van Buren & Mac Low 1992) and bow shocks observed ahead of runaway stars (Van Buren et al. 1995; Kaper et al. 1997). Observations carried out by the Einstein and ROSAT X-ray satellites have shown that high energy emission produced by the hot gas is ubiquitous in massive star forming regions, pervading the area occupied by the visible nebulosity (Chu & Mac Low 1990, Wang & Helfand 1991, Belloni & Mereghetti 1994; Norci & Oegelman 1995, Magnier et al. 1996; see also Chu 1994 and references therein). Modelling of the luminosity and energy distribution of high energy emission from star forming complexes shows that the combined action of supernovae and stellar winds can account for their X-ray emission (Chu & Mac Low 1990, Arthur & Henney 1996).
Actual star forming regions are far more complex than models can deal with: molecular clouds are highly inhomogeneous, their internal dynamics is greatly influenced by turbulent motions and magnetic fields, and massive stars tend to form in aggregates rather than in isolation. However, the champagne model is still useful in establishing a simple framework where the essential ingredients necessary to understand the dynamics of the interstellar medium around newly formed massive stars can be taken into account: large density discontinuities and the dominant energy sources from the central object. Therefore, both theoretical and observational evidence call for a reexamination of the champagne model, taking into consideration both photoionization and stellar winds (García-Segura & Franco 1996). This is the subject of this paper: we present the results of axisymmetric 2-D numerical simulations of a HII region expansion whose energy input includes both a stellar wind and photoionization. A comparison with photoionization-alone models is made.
We will focus on the large scale morphological and kinematic properties of the region, showing the distribution of the most relevant physical quantities. We have also produced simulated maps of X-ray emission in two broad bands, and maps of long wavelength recombination line to free-free continuum ratio, which can be useful for the interpretation of observations. Given the large range of geometrical and physical input parameters defining the scenario studied here, we do not intend to be exhaustive in exploring all the possible configurations; rather, we will concentrate on the dominant features of the blowout phase, leaving the application to specific sets of parameters to further studies of particular objects of interest.
In the next section, we describe the initial conditions assumed for the cloud and intercloud medium, as well as an outline of the numerical methods employed for simulating the expansion, and the relevant parameters of the exciting stars. Sect. 3 presents and describes the results obtained for the evolution of regions containing different stars at different distances from the cloud-intercloud interface, including the simulated maps at different wavelengths. Simulations without stellar wind or without thermal conduction are presented there as well for comparison. Our conclusions are summarized in Sect. 4.
© European Southern Observatory (ESO) 1997
Online publication: April 8, 1998