Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 358, 514-520 (2000)

Previous Section Next Section Title Page Table of Contents

1. Introduction

A new tool is now available to probe the population of recently formed massive stars, while still embedded in their parent clouds. Such stars are surrounded by a compact H II region, with an ionization front working outwards into the cloud. Regions undergoing massive star formation contain one or more ultra-compact H II (UCH II ) regions, and possibly more evolved H II regions. Bronfman et al. (1996, BNM) completed a survey in CS(2-1) towards IRAS point sources satisfying the Wood & Churchwell (1989a, WC89a) far-infrared (FIR) colour criteria for UCH II regions. The CS molecule is a tracer of high density molecular gas; a CS(2-1) detection strengthens the UCH II region identification and provides kinematic information. BNM detected 843 sources (hereafter IRAS/CS sources), whose azimuthally averaged galactic distribution is presented in Bronfman et al. (2000, BCMN). In this work we construct the luminosity function (LF) for the IRAS/CS sources, and show that it presents significant variations with galactocentric radius. We interpret the shape of the IRAS/CS sources LF and its variations in terms of an ensemble of massive star forming regions. The stellar content of IRAS/CS sources is characterised by the number of stars and their mass spectrum which, because of the youth of the systems, is taken to represent closely their initial mass function (IMF).

The use of young tracers to derive the IMF minimises the dependence on modelling, which afflicts most previous studies. The standard approach to determining the high mass end of the IMF has been through O and B star counts in the solar neighbourhood (e.g. Miller & Scalo 1979; Lequeux 1979). The average mass spectrum of newly formed stars is taken as a power law, and the value of the exponent characterises the IMF. These approaches require assumptions on the star formation history, and have the drawback of not probing the whole galactic disk. Garmany et al. (1982) addressed the question of large scale variations in the IMF exponent, and they favour a decrease with galactocentric radius (although their result has been reinterpreted, e.g. Massey 1998). A different tool was used by Vázquez & Feinstein (1989), who linked the variations in the open cluster luminosity function (Burki 1977) to variations in the IMF index.

Young objects such as H II regions have also been used to study massive stars in the galactic context. McKee & Williams (1997, MW97) characterised the population of newly formed massive stars through the luminosity function of OB associations. The ionising luminosity absorbed by the gas surrounding OB associations can be traced through the radio flux of H II regions, and the frequency distribution in luminosity of H II regions is in turn a function of the number of exciting stars and their masses. However, MW97 a-priori fixed the shape of the IMF and the distribution for the number of exciting stars. This method is also strongly model dependent due to the lack of complete information on galactic H II regions, and the fact that the population of H II regions is not homogeneous in age. Comerón & Torra (1996, CT96) also analysed the galactic disk distribution of newly formed massive stars, through a sample of IRAS point sources with colours of UCH II regions. But they did not use kinematic information to derive the galactic distribution of the UCH II regions, and their work is based on a position-independent LF. Wouterloot et al. (1995) used kinematic distances to build luminosity functions for samples of IRAS point sources with colours of young stellar objects. They studied the mass-spectrum of such sources either assuming a single exciting star, or defining an IMF-averaged luminosity with a variable upper mass limit.

The aims of this work are to present the FIR LF of massive star forming regions still embedded in their parent clouds; to investigate the LF large scale variations; and to analyse the LF in terms of the embedded stars' mass spectrum. In Sect. 2 we construct the LF of IRAS/CS sources for different sectors in the galactic disk, and show that it presents significant variations with galactocentric radius. In Sect. 3 we analyse the stellar population underlying the IRAS/CS sample through a simple model based on a synthetic ensemble of massive star forming regions (MSFRs). A search in parameter space is conducted in Sect. 4. The results of this analysis and the questions of the number and the mass spectrum of stars per MSFR will be addressed in Sect. 5. In particular, we will show that the mass spectrum of newly formed massive stars is constant across the galactic disk, although the average number of stars per MSFR is lower outside the solar circle. Sect. 6 is a brief estimate of the fraction of lifetime O stars spend in the embedded phase. In Sect. 7 we summarise our conclusions.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: June 8, 2000