The VLA-COSMOS Large Project is described and its scientific objective is discussed. We present a catalog of 3600 radio sources found in the 2deg^2^ COSMOS field at 1.4GHz. The observations in the VLA A and C configuration resulted in a resolution of 1.5"x1.4" and a mean rms noise of ~10.5(15)uJy/beam in the central 1(2)deg^2^. Eighty radio sources are clearly extended consisting of multiple components, and most of them appear to be double-lobed radio galaxies. The astrometry of the catalog has been thoroughly tested, and the uncertainty in the relative and absolute astrometry are 130 and <55mas, respectively.
Using the most recent (1998) version of the VLA FIRST survey radio catalog, we have searched for radio emission from 1704 quasars taken from the most recent (1993, Cat. <VII/158>) version of the Hewitt & Burbidge quasar catalog. These quasars lie in the roughly 5000 deg2 of sky already covered by the VLA FIRST survey. Our work has resulted in positive detection of radio emission from 389 quasars, of which 69 quasars have been detected for the first time at radio wavelengths. We find no evidence of a correlation between optical and radio luminosities for optically selected quasars. We find indications of a bimodal distribution of radio luminosity, even at a low flux limit of 1 mJy. We show that radio luminosity is a good discriminant between radio-loud and radio-quiet quasar populations, and that it may be inappropriate to make such a division on the basis of the radio-to-optical luminosity ratio. We discuss the dependence of the radio-loud fraction on optical luminosity and redshift.
We present flux densities and polarization percentages of 159 radio galaxies based on nearly simultaneous Very Large Array observations at four frequencies, 4.86, 8.46, 22.46, and 43.34GHz. This sample is selected from the high-frequency Australia Telescope 20GHz (AT20G) survey and consists of all sources with flux density S_20GHz_>40mJy in an equatorial field of the Atacama Cosmology Telescope (ACT) survey. For a subset of 25 of these sources, we used the Green Bank Telescope (GBT) to obtain 90GHz data. We find that, as expected, this sample consists of flatter spectrum and more compact or point-like sources than low-frequency-selected samples. In the K band, variability is typically <~20%, although there are exceptions. The higher frequency data are well suited to the detection of extreme gigahertz peak spectrum sources. The inclusion of the 43GHz data causes the relative fraction of inverted spectrum sources to go down and of peaked spectrum sources to go up when compared with the AT20G survey results. The trend largely continues with the inclusion of the 90GHz data, although ~10% of the sources with GBT data show a spectral upturn from 43GHz to 90GHz. The measured polarization fractions are typically <5%, although in some cases they are measured to be up to ~20%. For sources with detected polarized flux in all four bands, about 40% of the sample, the polarization fractions typically increase with frequency. This trend is stronger for steeper spectrum sources as well as for the lower flux density sources.
We present 1.4GHz catalogs for the cluster fields A370 and A2390 observed with the Very Large Array. These are two of the deepest radio images of cluster fields ever taken. The A370 image covers an area of 40'x40' with a synthesized beam of ~1.7" and a noise level of ~5.7uJy near the field center. The A2390 image covers an area of 34'x34' with a synthesized beam of ~1.4" and a noise level of ~5.6uJy near the field center. We catalog 200 redshifts for the A370 field. We construct differential number counts for the central regions (radius <16') of both clusters. We find that the faint (S_1.4GHz_<3mJy) counts of A370 are roughly consistent with the highest blank field number counts, while the faint number counts of A2390 are roughly consistent with the lowest blank field number counts. Our analyses indicate that the number counts are primarily from field radio galaxies. We suggest that the disagreement of our number counts can be largely attributed to cosmic variance.
We describe deep, new, wide-field radio continuum observations of the Great Observatories Origins Deep Survey-North field. The resulting map has a synthesized beam size of ~1.7" and an rms noise level of ~3.9uJy/beam near its center and ~8uJy/beam at 15' from phase center. We have cataloged 1230 discrete radio emitters, within a 40'x40' region, above a 5{sigma} detection threshold of ~20uJy at the field center. New techniques, pioneered by Owen & Morrison, have enabled us to achieve a dynamic range of 6800:1 in a field that has significantly strong confusing sources. We compare the 1.4GHz (20cm) source counts with those from other published radio surveys. Our differential counts are nearly Euclidean below 100uJy with a median source diameter of ~1.2". This adds to the evidence presented by Owen & Morrison that the natural confusion limit may lie near 1uJy. If the Euclidean slope of the counts continues down to the natural confusion limit as an extrapolation of our logN-logS, this indicates that the cutoff must be fairly sharp below 1uJy else the cosmic microwave background temperature would increase above 2.7K at 1.4GHz.
The Milky Way contains thousands of HII region candidates identified by their characteristic mid-infrared morphology, but lacking detections of ionized gas tracers such as radio continuum or radio recombination line emission. These targets thus remain unconfirmed as HII regions. With only ~2500 confirmed HII regions in the Milky Way, Galactic surveys are deficient by several thousand nebulae when compared to external galaxies with similar star formation rates. Using sensitive 9 GHz radio continuum observations with the Karl G. Jansky Very Large Array, we explore a sample of HII region candidates in order to set observational limits on the actual total population of Galactic HII regions. We target all infrared-identified "radio-quiet" sources from the Wide-field Infrared Survey Explorer Catalog of Galactic HII regions between 245{deg}>=l>=90{deg} with infrared diameters less than 80". We detect radio continuum emission from 50% of the targeted HII region candidates, providing strong evidence that most of the radio-quiet candidates are bona fide HII regions. We measure the peak and integrated radio flux densities and compare the inferred Lyman continuum fluxes using models of OB stars. We conclude that stars of approximately spectral type B2 and earlier are able to create HII regions with similar infrared and radio continuum morphologies as the more luminous HII regions created by O stars. From our 50% detection rate of "radio-quiet" sources, we set a lower limit of ~7000 for the HII region population of the Galaxy. Thus the vast majority of the Milky Way's HII regions remain to be discovered.
We present 33GHz imaging for 112 pointings toward galaxy nuclei and extranuclear star-forming regions at ~2" resolution using the Karl G. Jansky Very Large Array (VLA) as part of the Star Formation in Radio Survey. A comparison with 33GHz Robert C. Byrd Green Bank Telescope single-dish observations indicates that the interferometric VLA observations recover 78%+/-4% of the total flux density over 25" regions (~kpc scales) among all fields. On these scales, the emission being resolved out is most likely diffuse non-thermal synchrotron emission. Consequently, on the ~30-300pc scales sampled by our VLA observations, the bulk of the 33GHz emission is recovered and primarily powered by free-free emission from discrete HII regions, making it an excellent tracer of massive star formation. Of the 225 discrete regions used for aperture photometry, 162 are extranuclear (i.e., having galactocentric radii rG>=250pc) and detected at >3{sigma} significance at 33GHz and in H{alpha}. Assuming a typical 33GHz thermal fraction of 90%, the ratio of optically-thin 33GHz to uncorrected H{alpha} star formation rates indicates a median extinction value on ~30-300pc scales of A_H{alpha}_~1.26+/-0.09mag, with an associated median absolute deviation of 0.87mag. We find that 10% of these sources are "highly embedded" (i.e., A_H{alpha}_>~3.3mag), suggesting that on average, HII regions remain embedded for <~1Myr. Finally, we find the median 33GHz continuum-to-H{alpha} line flux ratio to be statistically larger within rG<250pc relative to the outer disk regions by a factor of 1.82+/-0.39, while the ratio of 33GHz to 24{mu}m flux densities is lower by a factor of 0.45+/-0.08, which may suggest increased extinction in the central regions.
We made two new sensitive (rms noise {sigma}_n_~1{mu}Jy/beam) high-resolution ({theta}=3.0" and {theta}=0.66" FWHM) S-band (2<{nu}<4GHz) images covering a single JVLA primary beam (FWHM~14') centered on RAJ2000=10:46,DEJ2000=+59:01 in the Lockman Hole. These images yielded a catalog of 792 radio sources, 97.7+/-0.8% of which have infrared counterparts stronger than S~2{mu}Jy at {lambda}=4.5{mu}m. About 91% of the radio sources found in our previously published, comparably sensitive low-resolution ({theta}=8" FWHM) image covering the same area were also detected at 0.66" resolution, so most radio sources with S(3GHz)>~5{mu}Jy have angular structure {phi}<~0.66". The ratios of peak brightness in the 0.66" and 3" images have a distribution indicating that most {mu}Jy radio sources are quite compact, with a median Gaussian angular diameter <{phi}>=0.3"+/-0.1" FWHM and an rms scatter {sigma}_{phi}_<~0.3" of individual sizes. Most of our {mu}Jy radio sources obey the tight far-infrared/radio correlation, indicating that they are powered by star formation. The median effective angular radius enclosing half the light emitted by an exponential disk is <{rho}_e_>~<{phi}>/2.43~0.12", so the median effective radius of star-forming galaxies at redshifts z~1 is <r_e_>~1.0kpc.
We have observed the Extended Chandra Deep Field-South (E-CDF-S) using a mosaic of six deep Very Large Array (VLA) pointings at 1.4GHz. In this paper, we present the survey strategy, description of the observations, and the first data release. The observations were performed during June through September of 2007 and included from 15 to 17 "classic" VLA antennas and 6 to 11 that had been retrofitted for the Expanded VLA (EVLA). The first data release consists of a 34.1'x34.1' image and the attendant source catalog. The image achieves an rms sensitivity of 6.4{mu}Jy per 2.8"x1.6" beam in its deepest regions, with a typical sensitivity of 8uJy. The catalog is conservative in that it only lists sources with peak flux densities greater than seven times the local rms noise, yet it still contains 464 sources. Nineteen of these are complex sources consisting of multiple components. Cross matching of the catalog to prior surveys of the E-CDF-S confirms the linearity of the flux density calibration, albeit with a slight possible offset (a few percent) in scale. Improvements to the data reduction and source catalog are ongoing, and we intend to produce a second data release in 2009 January.
Deep radio observations at 1.4GHz for the Extended Chandra Deep Field South were performed in 2007 June through September and presented in a first data release. The survey was made using six separate pointings of the Very Large Array with over 40hr of observation per pointing. In the current paper, we improve on the data reduction to produce a second data release (DR2) mosaic image. This DR2 image covers an area of about a third of a square degree, reaches a best rms sensitivity of 6uJy, and has a typical sensitivity of 7.4uJy per 2.8" by 1.6" beam. We also present a more comprehensive catalog, including sources down to peak flux densities of five or more times the local rms noise along with information on source sizes and relevant pointing data. We discuss in some detail the consideration of whether sources are resolved under the complication of a radio image created as a mosaic of separate pointings each suffering some degree of bandwidth smearing, and the accurate evaluation of the flux densities of such sources. Finally, the radio morphologies and optical/near-IR counterpart identifications are used to identify 17 likely multiple-component sources and arrive at a catalog of 883 radio sources, which is roughly double the number of sources contained in the first data release.
