Mid-infrared imaging and spectroscopy of the southern Hii region RCW 38

Craig H. Smith 2 Tyler L. Bourke 1 2 7 Christopher M. Wright 0 6 Henrik W. W. Spoon 6 David K. Aitken 5 Garry Robinson 2 John W. V. Storey 4 Takuya Fujiyoshi 2 Patrick F. Roche 3 Thomas Lehmann 8 0 Leiden Observatory, Postbus 9513, 2300 RA Leiden, the Netherlands 1 Harvard-Smithsonian Center for Astrophysics , 60 Garden Street, MS 42, Cambridge, MA 02138, USA 2 School of Physics, University College , ADFA, UNSW, Canberra, ACT 2600, Australia 3 Department of Astrophysics, Oxford University , Keble Road, Oxford OX1 3RH 4 School of Physics, The University of New South Wales , Sydney, NSW 2052, Australia 5 Division of Physical Sciences, University of Hertfordshire, College Lane , Hatfield, Herts AL10 9AB 6 Max-Planck-Institut fur extraterrestrische Physik , Postfach 1603, D-85740 Garching, Germany 7 AAO/ATNF Post-doctoral Fellow. Present address: Anglo-Australian Observatory, PO Box 296, Epping, NSW , 1710, Australia 8 Astrophysical Institute and University Observatory , Schillergachen 2, D-07745 Jena, Germany A B S T R A C T We present mid-infrared images and an 8-13 mm spectrum of the southern H ii region RCW 38. We determine the dust colour temperature from both our spectrum and images at 10 and 20 mm, and deduce the gas excitation from an image in the [S iv] fine-structure line, as well as spectra of the [Ar iii], [S iv] and [Ne ii] fine-structure lines. Our observations are consistent with a complex of sources associated with the RCW 38 IRS1 region, which represent knots of material in a shell, or ridge, surrounding a cavity of about 0.1 pc in radius, which is itself created by the stellar wind of the hot young source IRS2. The dust temperature does not peak closest to IRS2, but rather along the centre of the ridge, and is remarkably uniform over the extent of our image. From photoionization models for the observed line ratios at IRS1 we deduce a stellar effective temperature and gas density of about 43 000-48 000 K and 104 cm3 respectively. Whilst the star, or star cluster, IRS2 is ultimately responsible for the observed thermal and ionic emission, the relatively uniform dust temperature implies that the bulk of the dust heating in the region is provided by resonantly trapped Lyman a photons, rather than direct stellar photons. This then also implies that the dust is depleted with respect to the gas by a factor of at least 100 from its normal interstellar value. The small-scale spatial variations in the continuum emission and temperature can be explained by changes in the density and/or gas-to-dust mass ratio. I N T R O D U C T I O N RCW 38 (Rodgers, Campbell & Whiteoak 1960) is a luminous H ii region (, 7 105 L(; Furniss, Jennings & Moorwood 1975) located in the region of the Vela supernova remnant and the Gum nebula (Gum 1952, 1956). The reported distance to RCW 38 varies somewhat, but a reasonably consistent value is found amongst Radhakrishnan et al. (1972) of 1.5 kpc, Muzzio (1979) of 1.7 kpc, and Beck, Fisher & Smith (1991) of 1.760.9 kpc. In this paper we adopt a distance of 1.7 kpc for RCW 38. Frogel & Persson (1974, hereafter FP74) produced a 10-mm map of the region with a spatial resolution of 14.5 arcsec, showing the warm dust emission to be spatially extended in a horse-shoe shape over a 1.8 1.8 arcmin2 area. The peak of their map is labelled IRS1, but does not correspond with the bright 2.2-mm source IRS2 that dominates at near-infrared (NIR) wavelengths (FP74). FP74 found IRS1 to be extended on a scale of about 9 arcsec. Epchtein & Turon (1979) mapped a smaller (40 40 arcsec2) region with a spatial resolution of 7 arcsec. They found that the strong 10-mm peak IRS1 resolves into a number of more discrete sources, which they suggested may be due to local heating sources (i.e., a cluster of embedded young stars), or dust opacity variations. A low-resolution (warm filter) 10-mm spectrum of RCW 38 IRS1 is presented in Persson, Frogel & Aaronson (1976). From simple two-parameter fits to the spectrum they derive a dust temperature of 175 K and a modest absorption optical depth of t9:7mm 0:4. From its NIR colours FP74 suggest that IRS2 represents an early O star (around O4) or group of stars, suffering 12.8 magnitudes of extinction. As noted by Furniss et al. (1975), the ionizing flux from such a star is more than sufficient to account for all the 5-GHz flux observed by Goss & Shaver (1970). They suggested that an O5 star and little or no dust absorption of the continuum photons in the ionized region could fit the known parameters of RCW 38. Mizutani et al. (1987) present a Brg map of the region (15- arcsec resolution), and find that Brg emission largely follows the structure of the 10-mm images. From their observations they deduce an electron temperature of about 8000 K. They argue that IRS2 represents three O6 (ZAMS) stars rather than a single O5 star, although they do not preclude the possibility of an O5 star with a cluster of embedded stars. Storey & Bailey (1982) present a NIR K-band image of the inner 2.5 arcmin of RCW 38. Their image shows a number of discrete sources surrounded by extended nebulosity. NIR photometry indicates that all the bright sources in the field are highly reddened point sources, and they infer visual extinctions of up to 60 mag. Allen & Meadows (1992) present a JHK 0 image of a larger region around IRS2. Their more sensitive observations show that the nebulosity is far more extended than indicated by the image of Storey & Bailey. The blue appearance of the nebulosity in the JHK 0 false-colour image hints that the extinction toward the nebulosity is not high. Ligori et al. (1994) imaged a region , 1.5 1.5 arcmin2 at JHK centred on IRS2 with a spatial resolution of 0.89 arcsec pixel1. They resolve IRS2 into a cluster of at least five point sources within the central 16 arcsec, the beamsize used by FP74. The brightest source in this region dominates the observed flux and is saturated in their data. They also find a number of sources with large H K excess within their field of view, which may also be young embedded sources. One of their reddest sources is tentatively identified with FP74s IRS 3, detected at 10 mm. From their data Ligori et al. construct a K-band luminosity function, which they fit with a MillerScalo IMF to deduce a cluster age of , 2 Myr. Due to the complex nature of this source, the large-beam IRAS observations at 12 and 25 mm of this region do not provide much insight, although Kuiper et al. (1987) are able to extract useful 60and 100-mm fluxes, from which they deduce a cool dust temperature of 45 K and gas column density of 2 1020 cm2. However, from observations at 1 mm Cheung et al. (1980) derive a high column density of 8 1023 cm2, from which they infer a large visual extinction of 800 mag, indicating that RCW 38 lies at the front surface of a dense and massive (104 M() molecular cloud. Detailed molecular observations of RCW 38 are rare. Gillespie, White & Watt (1979) mapped a region 1.0 0.5 (...truncated)