Title Follow-up observations of high-redshift submillimeter galaxies Pi A. Blain Time 5300 hrs 1. Name of program and authors Follow-up observations of high-redshift submillimeter galaxies Andrew Blain 2. One short paragraph with science goal(s) Since the first sensitive submm-wave surveys in 1997, a steadily increasing sample of high-redshift galaxies have been discovered, with 300-GHz flux densities of order 5mJy. There are currently about 600 of these galaxies known, and about 120 have redshifts mostly in the interval z=2-3. By the time ALMA operations begin, it is possible that 10,000 such galaxies may be known, with perhaps 1-2,000 redshifts. This depends on the success of single-antenna telescopes like JCMT/SCUBA-2, APEX and the 50-m LMT with wide-field bolometer array detectors. A handful have high-resolution mm-wave interferometer detections, showing emission resolved typically on the sub-arcsec scale, but there are so far no measurements close to the peak of the SED of these galaxies at about 100 microns. ALMA will provide multi-color spectral images of these galaxies longward of their SED peak to probe their astrophysics. Resolved images of these galaxies will provide valuable information about the relationship between optical and mm/submm morphology for the first time, providing details about the reprocessing of light in these galaxies, the temperature distribution of the dust, etc. Where redshifts are known, receiver tunings will be chosen to include CO lines in the ALMA bands if possible. 3. Number of sources (e.g., 1 deep field of 4'x4', 50 YSO's, 300 T Tauri stars with disks, ...; do NOT list individual sources or your "pet object", except in special cases like LMC, Cen A, HDFS) Of order 1000-2000 galaxies, spread over the sky. Many fields are currently in the North. Future fields will be equatorial, including COSMOS and follow-up of VLT-VIMOS redshift survey. This large number can be used to construct an accurate luminosity function, spanning the range of luminosities from typical galaxies to the most extreme systems, and rooting out the effects of gravitational lensing etc. However, some useful statistical information is available from a sample size as small as 300 targets, allowing a 10-bin luminosity function to be compiled. 4. Coordinates: 4.1. Rough RA and DEC (e.g., 30 sources in Taurus, 30 in Oph, 20 in Cha, 30 in Lupus) Indicate if there is significant clustering in a particular RA/DEC range (e.g., if objects in one particular RA range take 90% of the time) Should be approximately uniform. 10hr equatorial field, 02hr southern fields are promising targets for APEX and LMT surveys accessible to ALMA. Could be bunching in the 02hr region, based on GOODS-S field. 4.2. Moving target: yes/no (e.g. comet, planet, ...) No 4.3. Time critical: yes/no (e.g. SN, GRB, ...) No 4.4. Scheduling constraints: (optional) Good weather required, 02hr and 10hr likely to most popular RA. 5. Spatial scales: 5.1. Angular resolution (arcsec): 0.01"-1" 5.2. Range of spatial scales/FOV (arcsec): (optional: indicate whether single-field, small mosaic, wide-field mosaic...) Single field in general, targeted at known object. 5.3. Required pointing accuracy: (arcsec) 1" 6. Observational setup 6.1. Single dish total power data: no/beneficial/required No Observing modes for single dish total power: (e.g., nutator switch; frequency switch; position switch; on-the-fly mapping; and combinations of the above) 6.2. Stand-alone ACA: no/beneficial/required No, unless standing idle for a bit more collecting area. 6.3. Cross-correlation of 7m ACA and 12m baseline-ALMA antennas: no/beneficial/required Yes, if ACA used 6.4. Subarrays of 12m baseline-ALMA antennas: yes/no No 7. Frequencies: 7.1. Receiver band: Band 3, 4, 5, 6, 7, 8, or 9 In principle all bands to provide excellent SED, but 3,6 and 9 should provide good continuum SED, with the best chance of detecting a CO line coming in 3, 4 & 6 where fractional bandwidth is greatest. Where CO/HCN will fall in band, at known redshift, want to observe there. 7.2. Lines and Frequencies (GHz): (approximate; do _not_ go into detail of correlator set-up but indicate whether multi-line or single line; apply redshift correction yourself; for multi-line observations in a single band requiring different frequency settings, indicate e.g. "3 frequency settings in Band 7" without specifying each frequency (or give dummies: 340., 350., 360. GHz). For projects of high-z sources with a range of redshifts, specify, e.g., "6 frequency settings in Band 3". Apply redshift correction yourself.) Single setting in each band. Follow up may require more complex setups to hunt more unusual lines. 7.3. Spectral resolution (km/s): 100-300 km/s 7.4. Bandwidth or spectral coverage (km/s or GHz): 8GHz, max continuum sensitivity. 8. Continuum flux density: 8.1. Typical value (Jy): (take average value of set of objects) (optional: provide range of fluxes for set of objects) For typical galaxy at z=2.5 or so: 90GHz ~0.08mJy 230GHz ~2mJy 350GHz ~5mJy 670GHz ~15mJy 8.2. Required continuum rms (Jy or K): Need to obtain high signal-to-noise resolved images to determine= detailed morphologies: implies at least 10-sigma detections: The most sensitive band in terms of signal to noise is Band 7, the highest resolution in any configuraion is expected in Band 9; therefore, the signal-to-noise ratio targets in these bands should be the greatest. 90GHz 5microJy 230GHz 0.05mJy 350GHz 0.05mJy 670GHz 0.1mJy 8.3. Dynamic range within image: (from 7.1 and 7.2, but also indicate whether, e.g., weak objects next to bright objects) Small. 100-1000 8.4. Calibration requirements: absolute ( 1-3% / 5% / 10% / n/a ) repeatability ( 1-3% / 5% / 10% / n/a ) relative ( 1-3% / 5% / 10% / n/a ) Absolute 10%, relative band-to-band would like 5% for accurate SEDs. 9. Line intensity: Difficult to be sure, but in Band 3 there is a good chance that CO(3-2) can be detected for most sources, revealing internal dynamics and relative distributions of gas, dust and stars. 9.1. Typical value (K or Jy): (take average value of set of objects) (optional: provide range of values for set of objects) Few mJy over 300 km/s channel 9.2. Required rms per channel (K or Jy): See above, but set by continuum 9.3. Spectral dynamic range: N/A 9.4. Calibration requirements: absolute ( 1-3% / 5% / 10% / n/a ) repeatability ( 1-3% / 5% / 10% / n/a ) relative ( 1-3% / 5% / 10% / n/a ) 10. Polarization: yes/no (optional) No 10.1. Required Stokes parameters: 10.2. Total polarized flux density (Jy): 10.3. Required polarization rms and/or dynamic range: 10.4. Polarization fidelity: 10.5. Required calibration accuracy: 11. Integration time for each observing mode/receiver setting (hr): 90GHz 100min 230GHz 6min (at 280GHz only 10min for better S/N). 350GHz 367min 670GHz 167min => 640 minutes each 12. Total integration time for program (hr): 500 sources (estimated, most likely with redshifts) => 5300 hours total 13. Comments on observing strategy : (optional) (e.g. line surveys, Target of Opportunity, Sun, ...): Note that there are a wide variety of other point source targets for ALMA: optically-selected high-redshift Lyman-break galaxies (with ~0.1mJy at 350GHz); near-infrared-selected ERO galaxies (some of which are several mJy at 350GHz); galaxies detected by SIRTF (maybe 10,000,000 in the catalog). A key extra sample of galaxies are those detected by Planck Surveyor. There are likely to be several thousand of these (see additional proposal). I include an extra Planck Surveyor proposal, and I know there are high-redshift QSO proposals. I can certainly foresee at least comparable amounts of time being required to survey the field. One option to reduce time greatly would be to make a single survey at 230 GHz or 280 GHz first, and then to sift the results for follow up. This could reduce the time required by a factor of 10. -------------------------------------------------- Review v2.0: 1.1.8 I fully agree with the author's proposal; making a single band survey at 280 GHz first (500 sources * 10min for 0.05 mJy rms, i.e., 83 hrs), given the huge amount of requested time. Note that the sensitivity of 0.05 mJy rms at 350 GHz will be accomplished by about 30 min integration, not 367 min. Thus, proposed observations will require (100+6+30+167min) * 500 sources = 2500 hrs in total.