2 G29.96-0.02: Observational facts

The compact H II region G29.96 is one of the brightest radio and infrared sources in our Galaxy. Its morphology is a classical example of a cometary-like compact H II region (Wood & Churchwell 1989b) in interaction with a molecular cloud (e.g., Pratap et al. 1999, hereafter PMB99). G29.96 has been studied in detail at infrared and radio wavelengths and the following summarizes the main results together with the essential properties of this compact H II region.

     
2.1 Observations of fine-structure lines

The fine-structure lines observed by the SWS and LWS spectrometers are tabulated in Paper I where their observed fluxes associated error bars are given as well as the entire ISO spectrum of this source. In addition, 11 hydrogen recombination lines were detected in G29.96 and were used to determine the interstellar extinction (Paper II). We will use Br$\alpha$ in the modeling process (see Sect. 4.1 and Table 1) as this is the brightest H I line and one of the least affected by extinction. The ISO lines corrected from the interstellar extinction (using A_K=1.6 and the extinction law derived in Paper II) are given in Col. 2 of Table 2.

Infrared fine-structure lines have previously been observed with KAO by Herter et al. (1981), who measured the lines [Ar II] 6.98$~\mu$m, [Ar III] 8.98$~\mu$m, [S III] 18.7$~\mu$m and [Ne II] 12.8$~\mu$m, in agreement with the ISO values within 20%.

Similarly, Simpson et al. (1995) reported observations of [S III] 33.6$~\mu$m, [Ne III] 36.0$~\mu$m, [O III] 51.8$~\mu$m, [N III] 57.3$~\mu$m and [O III] 88.3$~\mu$m differing from the ISO values by less than 10%, except the [S III] 33.6$~\mu$m and the [O III] 88.3$~\mu$m fluxes which are both $\sim$30% higher in our sample. The difference of aperture sizes and pointing between KAO and ISO can partly be responsible of these discrepancies.

Maps of [Ne II] 12.8$~\mu$m were obtained by Lacy et al. (1982) and Watarai et al. (1998). These authors found an integrated flux of 84% and 64% of our value, over a size of 10 $^{\prime\prime}$$\times$10 $^{\prime\prime}$ and a diameter of 30 $^{\prime\prime}$ respectively.

    
2.2 Radio observations: Core/halo structure and constraints on the He ionization structure

G29.96 has been observed at radio frequencies using various spatial resolutions. At 2 cm the observed flux densities range from 2.7 to 4.6 Jy (Wood & Churchwell 1989b; Afflerbach et al. 1994; Fey et al. 1995; Watson et al. 1997), at 6 cm from 1.4 to 3.6 Jy (Wood & Churchwell 1989b; Afflerbach et al. 1994), and at 21 cm from 0.9 to 2.6 Jy (Claussen & Hofner 1995; Kim & Koo 2001). As usually the case for UCHII regions, the source presents a diffuse emission. Therefore the highest spatial resolution observations may miss part of the radio flux density. We have adopted the following values: 3.9, 3.4 and 2.6 Jy at 2, 6, and 21 cm respectively, favoring the highest values to take into account the diffuse emission. The resulting number of Lyman continuum photons ( $1.8 \times 10^{49}$, cf. below) is higher than the value of $4 \times 10^{48}$derived in Paper II, but this last value was obtained for a uniform electron density of 10⁴ cm^-3 and a size of 7 arcsec.

G29.96 is characterized by a strong edge-brightened core/"head'', with a low surface-brightness "tail'' of emission trailing off opposite the bright edge. Due to this extended emission, part of G29.96 cannot be strictly called Ultra Compact. Wood & Churchwell (1989b) obtained $n_{\rm e} = 8.5\times 10^4$ cm^-3 in the arc and a $n_{\rm e}$ 5-10 times lower in the tail of the nebula, while Afflerbach et al. (1994) estimate $n_{\rm e} = 5.\times 10^4$ cm^-3 in the leading arc and $n_{\rm e} = 2.\times 10^4$ cm^-3 in the tail. Simpson et al. (1995) obtained $n_{\rm e} = 1500~{\rm ~cm}^{-3}$ from the [O III] 51.8/88.3$~\mu$m lines ratio. From the set of the ISO observables available for G29.96, 3 line ratios can be used as density diagnostics: [O III] 51.8/88.3$~\mu$m, [S III] 18.7/33.6$~\mu$m and [Ne III] 15.5/36.0$~\mu$m. Unfortunately, the two last ratios suffer large calibration uncertainties (25% at 1 $\sigma$ error, see Paper I) and therefore only [O III] 51.8/88.3$~\mu$m (hereafter $r_{[O {\sc iii}]}$) can be safely used to derive the gas density. From $r_{[O {\sc iii}]}$ = 2.4 an electron density $n_{\rm e}$ $\sim$ 800 cm^-3 is derived (Paper II).

This apparent discrepancy between the densities determined from fine-structure lines of oxygen and from radio observations have already been pointed out by Afflerbach et al. (1997) who suggested a core/halo description of G29.96. Faint diffuse halos are commonly observed in the radio continuum maps of UCHII regions (e.g., Garay et al. 1993; Fey et al. 1995; Afflerbach et al. 1996; Kurtz et al. 1999). Recently, Kim & Koo (2001) found extended emission at 21 cm linked to the bright spot of G29.96. Br$\alpha$ images (Lumsden & Hoare 1996; Watson et al. 1997; Watson & Hanson 1997, PMB99) also support this morphology.

The observational evidence of a core/halo morphology has led us to model G29.96 with two components as outlined in Sect. 3.3.

Kim & Koo (2001) have observed the He76$\alpha$ and H radio recombination lines in G29.96. For various positions, including the region considered here, they obtain a He⁺ abundance of $\sim$0.068-0.076. Assuming a normal helium abundance of 0.1 this implies that He is predominantly singly ionized helium in G29.96. This provides a useful constraint on the temperature of the ionizing source (see Sect. 5.7).

   
2.3 Heliocentric and galactocentric distances

With an average global radio recombination lines LSR velocity of 95 kms^-1 (Afflerbach et al. 1994), adopting a galactocentric radius of the Sun of 8.5 kpc, a kinematic heliocentric distance for G29.96 between 5.5 and 9.5 kpc is derived using a standard galactic rotation curve with a rotation speed of 220 kms^-1 at the Suns position. Previous investigators assumed the average between the near and far distances.

The extinction along the line of sight to G29.96 was studied by PMB99 using galactic H I and CO surveys. They found $A_K\sim 1$ at a distance of 5 kpc, and $A_K\sim 3$ at a position corresponding to the tangent point (at about 7.5 kpc). These findings provide a crucial argument to consider that the near distance should be more appropriate. Hereafter, we will adopt an heliocentric distance of 6 kpc (+1.0, -0.5) for G29.96, implying $R_{\rm G}=4.5$ kpc (+/-0.3).

   
2.4 Ionizing star(s)

A detailed search for stars embedded in the H II region and the adjacent molecular hot core has been performed in the near-infrared (Lumsden & Hoare 1996; Watson et al. 1997; Watson & Hanson 1997, PMB99). Watson et al. (1997) and PMB99 have in particular revealed the existence of a cluster of about 18 OB-type stars, or their progenitors, embedded in the cloud. The same authors have convincingly identified the bright star at the center of the arc of radio emission as the exciting star, or at least as the primary source of ionization. There is no evidence for an infrared excess in the K band, suggesting that any remaining disk should be optically thin, and therefore that the star is no longer accreting mass. From near infrared spectroscopic observations, Watson & Hanson (1997) were able to constrain the spectral type of the ionizing source. A spectral type between O5 and O8 (and no constraint on the luminosity class) was found. In contrast, a recent study by Kaper et al. (2002) reports a spectral type as early as O3. From their spectroscopic and photometric data Watson & Hanson (1997) and Watson et al. (1997) already deduced an evolutionary age of about $1{-}2\times 10^6$ years for a single or binary star, in apparent contradiction with the estimated age of the UCHII region ($\sim$10⁵ yr).

The overall SED of G29.96 derived by Watson et al. (1997), the near-IR photometry, and spectral types provide important constraints on the fundamental properties ( $T_{{\rm eff}}$, luminosity, age) of the ionizing source. A detailed discussion is given in Sect. 5.7.