Title Unbiased survey of submm galaxies: combined line+cont at 3 mm Pi S.Guilloteau Time 256 hrs 1.1.2: Name -- Unbiased survey of submm galaxies -- -- Part 2 -- Authors: S.Guilloteau 2. Science goal: -- (Generic Science Scope from C.Carilli) -- The discovery of the IR background, and the SCUBA/MAMBO population of dusty, star forming galaxies at high redshift, has transformed our understanding of galaxy formation. It is now clear that a significant fraction (of order 50%) of star formation in the cosmos occurs in galaxies that are heavily obscured by dust, and that this fraction may rise with redshift, possibly corresponding to the formation of spheroidal galaxies in active starbursts. One highly uncertain aspect of the study of submm galaxies is their redshift distribution. Optical redshifts remain problematic for the majority of such sources, and can be misleading due to possible mis-identifications. -- (specific part) An important way to study the history of galaxy formation is to perform an unbiased redshift survey. We propose a unbiased, high sensitivity survey over a 4x4' area of the sky. This survey consists in a 4 main parts, and 2 or 3 complementary observations which are discussed in separate proposals: - Part 1) Continuum survey at 1 mm to reach 0.1 mJy point source sensitivity at the 5 sigma level. - Part 2) Combined line and continuum survey at 3 mm, down to 7.5 microJy at the 5 sigma level. - Part 3) Pointed continuum survey at 650 GHz towards sources detected in Part 1), down to 0.4 mJy at the 5 sigma level - Part 4) Line survey of the 210-275 GHz frequency band, down to 50 microJy for the 8 GHz bandwidth, with some angular resolution (0.4") in order to provide dynamical masses and lens corrections for the detectable sources. The fraction of sources found in Part 1 to be detected in Part 4 is unknown, but should be high (perhaps 50 -- 80 %). The complementary observations include: higher resolution imaging for better lens models, and better derivation of dynamical masses, and observations of other lines such as HCN or CI. This proposal covers Part 2). In Part 2), we propose to cover completely in space the same area as covered in Part 1), and in frequency the range 84 to 116 GHz. The frequency coverage can be done in 4 tunings of the Band 3 receivers, although the details depend on the final choice of receivers for Band 3. We propose to achieve a continuum sensitivity (averaged over the total 32 GHz) of 7.5 microJy at 5 sigma all over the 4x4' field. Such a field requires about 16 pointings (slightly less at 84 GHz, slightly more at 116 GHz). At 90 GHz, an integration time of 12 hours per pointing is required to reach this sensitivity level, leading to a total time of 8 days to perform the program. Given the degraded sensitivity at 84 GHz (receiver) and 115 GHz (atmospheric oxygen), Part 2 can be completed in about 10 days. With a typical spectral index of 2-3 for the dust emission between 3 mm and 1 mm, depending on the source redshift, sources detected in Part 1 at the 0.2 mJy level will have a 3 mm continuum flux between 8 and 23 microJy. The continuum sensitivity is thus sufficient to allow detection of all these sources at 3 mm as well, and to provide a first order information on the redshift by the determination of the spectral index. For a spectral resolution of 50 km/s, the corresponding line sensitivity is 70 microJy (taking into account the 4 times smaller integration time for each receiver tuning). For typical linewidths of 300 km/s, the integrated line flux sensitivity is thus 0.04 Jy.km/s (at 5 sigma). Current (as of July 2003) detection of CO in sub-mm galaxies indicate an integrated line flux at 3 mm of order 1 to 2 Jy.km/s for sources of about 5 to 10 mJy continuum flux at 1 mm. As for the 3 mm continuum, the 5 sigma sensitivity is sufficient to detect in CO all continuum sources detected above 0.2 mJy in Part 1). In detected sources, there will be at least one CO line for sources with z>2, and two for sources with z>6. The are two blind redshift regions: 0.4 -- 1, and 1.7 -- 2.0. The survey thus provides a good coverage of the star formation history for redshifts above 1. 3. Number of sources: 1 field of 4x4' 4. Coordinates: 4.1. Any 4.2. Moving target: no 4.3. Time critical: no 5. Spatial scales: 5.1. Angular resolution: 3" 5.2. Range of spatial scales/FOV: scales < 1", FOV 4x4' 5.3. Single dish: no 5.4. ACA: no 5.5. Subarrays: no 6. Frequencies: 6.1. Receiver band: Band 3 -- 84 -- 116 GHz 6.2. Lines and Frequencies 4 tunings for complete band coverage 6.3. Spectral Resolution (km/s) 50 6.4. Bandwidth or spectral coverage: 8 GHz per tuning 7. Continuum flux density: 7.1. Typical value: 0.01 - 0.1 mJy 7.2. Continuum peak value: 0.2 mJy 7.3. Required continuum rms: 1.5 microJy 7.4. Dynamic range in image: < 100:1 8. Line intensity: 8.1. Typical value: 8.2. Required rms per channel: 14 microJy 8.3. Spectral dynamic range: < 30:1 9. Polarization: no 10. Integration time per setting: Mosaic of 16 pointings, 4 hours per pointing center and per tuning, 4 tunings. 11. Total integration time for program: Part 2): 256 hours ************************************************************************ Review Pierre Cox: this programme in four parts proposes to explore a 4x4 arcmin^2 field in order to study the submm galaxy population in both the dust and line emission. A few 100-300 sources will be detected providing a census of the population together with first indications on their properties (redshifts, infrared luminosities, gas excitation, and dynamical masses). The time estimates are correct and the overall strategy is robust. Note that in 1.1.2, the current CO detections of submm galaxies indicate that the CO fluxes are about 1-2 Jy km/s for 1.3 mm continuum flux densities of 1-3 mJy rather than the 5-10 mJy which are indicated and which correspond to the 850 microns flux densities. Also at the expected sensitivities, on could perhaps start to detect species other than CO such as HCN which is typically 10 times weaker than CO. -------------------------------------------------- Review v2.0: 1.1.2 Time estimation was checked. Note that "14 microJy" stated in 8.2. Required rms per channel seems to be 70 microJy.