4.1. The Orion nebula: a benchmark
The Orion nebula is the brightest and most observed H II region in the galaxy. Therefore it is a benchmark in many respects. O'Dell (2001) and Ferland (2001) have summarized our present knowledge on this object. In the following, we only discuss aspects related to the chemical composition in the ionized gas.
It is of interest, beforehand, to mention that it is with the Orion nebula that the concept of filling factor started. Using a spherical representation, Osterbrock & Flather (1959) showed that the optical surface brightness data could be reconciled with the observed [O II] 3726/3729 intensity ratios only when assuming extreme density fluctuations. They proposed a schematic model in which these fluctuations are represented as condensations immersed in a vacuum, with the relative volume of the condensations being only 1/30 of the total volume of the nebula. But a more realistic model of the Orion nebula (Zuckerman 1973, Balick et al. 1974, see also discussion in O'Dell 2001) is to represent the Orion nebula as an ionized blister on a background molecular cloud. From a detailed comparison of the H surface brightness map and of the [S II] 6731/6717 map, Wen & O'Dell (1995) derived a 3D representation of the nebula. The ionized skin is very thin with respect to the overall size of the nebula, which justifies the plane parallel approximation for photoionization modelling.
The extinction in the Orion nebula is well known to differ from the standard reddening law, and has been studied in detail (see Baldwin et al. 1991, Bautista et al. 1995, Henney 1998 for recent references).
Abundances have been derived both from Te-based empirical methods and from photoionization models, using optical data with increased signal to noise and spectral resolution, with the addition of ultraviolet data obtained with IUE (and more recently with HST) and infrared data from ground-based telescopes and from KAO, ISO, and MSX. Table 2 summarizes the abundances derived during the last decade. All the abundances are given in ppM units (106 × the number of particles of a given species with respect to hydrogen) .
|a Rubin et al. (1991, 1993), optical + IR spectroscopy, model|
|b Baldwin et al. (1991), long slit optical + IR + UV spectroscopy, model|
|c Osterbrock et al. (1992), optical spectroscopy, empirical|
|d Esteban et al. (1998) (t2 = 0.024), optical spectroscopy, empirical|
|e Esteban et al. (1998) (t2 = 0.0), optical spectroscopy, empirical|
|f Simpson et al. (1998), IR spectroscopy, empirical|
|g Deharveng et al. (2000), optical integrated spectroscopy, empirical|
There is rather good agreement for the oxygen abundances, the value of Esteban et al. (1998) with t2 = 0 being the lowest and the one with t2 = 0.024 being the highest. Note that the preferred abundances of Esteban et al. (1998) are those obtained with t2 = 0.024, which is the value indicated by the ORL/CEL comparison. However, the comparison of the [O III] 4363/5007 and Balmer jump temperatures is consistent with t2 = 0. One must be aware that abundances from models are not always the most reliable, since the models do not reproduce the ionization structure perfectly. The values of Simpson et al. (1998) for Ne, S, and Ar are obtained from simultaneous observations of the most abundant ionic stages.
The Mg, Si, Fe and Ni abundances are heavily depleted with respect to the Sun (indicating the presence of grains intimately mixed with the gas phase in the ionized region). There is actually a controversy with respect to the interpretation of Fe lines (Bautista et al. 1994, Baldwin et al. 1996, Bautista & Pradhan 1998). Esteban et al. (1999) recommend to derive Fe abundances from Fe++ lines as done in the works quoted in Table 2.