The APEX southern sky survey of high mass star forming regions

During my dissertation work, I studied the early stages of massive star formation in the southern hemisphere, using sub-mm line and continuum data. A sample of 47 sources, chosen according to IRAS color selection criteria, was selected. The properties of this sample have been characterized both in line and continuum emission and it can serve as a suitable sample for follow-up observations with ALMA.Pilot study: I first conducted a pilot study on three very luminous, line rich sources out of the sample, which also served to test and refine the selection of frequency setups to be used for the large sample. The three sources were observed in line emission in a number of frequency setups between 290 GHz and 810 GHz to determine their physical and chemical properties, as well as in 870 µm dust continuum with the Large APEX BOlometer CAmera (LABOCA). The three sources were further observed at high resolution mm continuum and line emission with the Australia Compact Telescope Array (ATCA). They are all luminous, massive objects with signs of outflow activity, are rich in sulfur-bearing species and display a high abundance of oxygen- versus nitrogen-bearing species. One of them seems to have a highly complex velocity structure. Altogether, the three sources show evidence of being hot molecular cores at a more evolved stage.High mass star formation sample - continuum: To characterize the sample of 47 sources, they were studied in 870 µm dust continuum with LABOCA and their morphologies and associations with mid-infrared sources were investigated. About 80% of the sources have an infrared counterpart within 0.1 pc of the dust peak, indicating luminous heating sources. Radial fits to the data resulted in an average density profile of ~r-1.6, which is consistent with the picture of proto-stellar density power law distribution (Shu et al., 1987, ARA&A, 25). Comparison of the subsamples with (radio-loud) and without (radio-quiet) detected radio continuum shows that the radio-loud sources are on average more massive, luminous and have higher pressures. The luminosity-to-mass ratio L/M found in our sample shows no trends with mass, consistent with a constant star forming rate in the Galaxy (Shirley et al., 2008, ASPC, 387). I compared the sources in the sample with a model of high mass proto-stellar evolution (Molinari et al., 2008, A&A, 481). This comparison suggests that both radio-loud and radio-quiet sources have finished the accelerated accretion phase after about 2x105yr.High mass star formation sample - molecular line survey: A subsample of the 47 sources has also been observed with APEX in a partial line survey around 338 GHz, allowing me to characterize the chemical properties of the sample. Based on their spectra, the sources could be classified as either line poor or line rich (Hatchell, 1998, A&S, 133). For most sources, only contribution from a cool, extended envelope was detected, while some of the line-rich sources show also a hot, compact component. It turned out that the overall chemical make-up of the envelopes in the sources is fairly similar, despite their wide range of masses and luminosities. Analyzing infall properties of the sample, a statistical trend towards infalling motions could be found. I found eight hot cores among the sample, all of which show a similar trend towards being oxygen-rich. Together with their 870 µm dust continuum properties and archival cm continuum data, I tentatively placed the hot cores in our sample in a transition stage between hot cores and UCHII regions. Having found various sources which have the same mass and luminosity properties as the hot cores, but do not exhibit hot core type molecular line spectra, the question is raised why only some of the sources develop hot cores. Using the line observations as a chemical magnifying glass to zoom into the deeply embedded inner regions of the hot cores, I estimated mass accretion rate, star forming time and the pressures of the star forming core in the frame of the turbulent core model (McKee & Tan, 2003, ApJ, 585)