📝 Original Info
- Title: The Deep SWIRE Field II. 90cm Continuum Observations and 20cm-90cm Spectra
- ArXiv ID: 0904.2011
- Date: 2009-04-13
- Authors: Frazer N. Owen, G. E. Morrison, Matthew D. Klimek, Eric W. Greisen
📝 Abstract
We present one of the deepest radio continuum surveys to date at a wavelength ~1 meter, at 324.5 MHz. The data reduction and analysis are described and an electronic catalog of the sources detected above 5 sigma is presented. We also discuss the observed angular size distribution for the sample. Using our deeper 20cm survey of the same field, we calculate spectral indices for sources detected in both surveys. The spectral indices for 90cm-selected sources, defined as S ~nu^(-alpha}, shows a peak near 0.7 and only a few sources with very steep spectra. Thus no large population of very steep spectrum microJy sources seems to exist down to the limit of our survey. For 20cm-selected sources, we find similar mean spectral indices for sources with S_20>1 mJy. For weaker sources, below the detection limit for individual sources at 90cm, we use stacking to study the radio spectra. We find that the spectral indices of small (<3") 20cm-selected sources with S_20< 10 mJy have mean and median alpha(90,20)~0.3-0.5. This is flatter than the spectral indices of the stronger source population. We report log N-log S counts at 90cm which show a flattening below 5 mJy. Given the median redshift of the population, z~1, the spectral flattening and the flattening of the log N-log S counts occurs at radio luminosities normally associated with AGN rather than with galaxies dominated by star-formation.
💡 Deep Analysis
Deep Dive into The Deep SWIRE Field II. 90cm Continuum Observations and 20cm-90cm Spectra.
We present one of the deepest radio continuum surveys to date at a wavelength ~1 meter, at 324.5 MHz. The data reduction and analysis are described and an electronic catalog of the sources detected above 5 sigma is presented. We also discuss the observed angular size distribution for the sample. Using our deeper 20cm survey of the same field, we calculate spectral indices for sources detected in both surveys. The spectral indices for 90cm-selected sources, defined as S ~nu^(-alpha}, shows a peak near 0.7 and only a few sources with very steep spectra. Thus no large population of very steep spectrum microJy sources seems to exist down to the limit of our survey. For 20cm-selected sources, we find similar mean spectral indices for sources with S_20>1 mJy. For weaker sources, below the detection limit for individual sources at 90cm, we use stacking to study the radio spectra. We find that the spectral indices of small (<3") 20cm-selected sources with S_20< 10 mJy have mean and median al
📄 Full Content
We are building a deep multi-wavelength picture of the sky in the SWIRE Spitzer deep field, 1046+59, which was chosen to be ideal for deep radio imaging. In paper I we discussed the 20cm continuum survey . The present 90cm survey allows us to study the radio spectra of the general source population. For Jansky and mJy sources, very steep radio spectra often are associated with very high redshifts, although the physical origin of this effect remains unclear (e.g., Miley & De Breuck 2008). A large population of very steep spectrum, µJy sources might suggest a corresponding high redshift µJy population. On the other hand, flatter radio spectra are often thought to be connected with synchrotron self-absorption or free-free absorption, although other mechanisms could potentially produce such spectra. Combined with other information the low frequency spectral energy distribution has the potential to give us unique insight on the physics of black-hole-driven AGN and star-forming galaxies. In this paper we report our 90cm observations with the VLA and some analysis of these radio data combined with our 20cm survey of the same field from paper I. In future papers in this series, we will combine these data with redshift measurements and observations at other wavelengths.
Observations were made of a single pointing center position, 10 h 46 m 00 s , 59 • 01 ′ 00 ′′ (J2000), with the VLA in A and C configurations for a total of almost 85 hours on-source between February 2006 and January 2007. However, due to the ongoing EVLA upgrade, only 22 working antennas were typically available in A and 18 in C. Thus the total integration time was equivalent to ∼63 hours in A and ∼5 in C, with correspondingly less uv coverage.
In Table 1, we summarize the parameters of the observing runs. Since the total time is dominated by the A configuration, the final image for analysis had a resolution ∼ 6 ′′ and FWHM FOV of 2.3 • . The data were all taken in spectral-line mode 4 using on-line Hanning smoothing, resulting in fifteen 390.625 kHz channels in each of 2 IFs (centered at 321.5 and 327.5 MHz) and each of two polarizations. Five second integration times were used in the A configuration and 10 seconds in C. The integration times and channel bandwidths were chosen to minimize tangential and radial smearing of the images away from the field center. This combination of parameters produces the best compromise for imaging sensitivity and quality possible with the current VLA correlator, which dates from the 1970’s. The finite bandwidth of the spectral channels still produces some radial smearing of the image away from the field center which we take into account in the analysis of the image.
For calibration, editing, and imaging a procedure similar to the one described in paper I was used. The Baars flux density scale (Baars et al. 1977) was adopted using 3C286 as the flux calibrator. Two of the 15 channels in each IF were deleted due to interference which is generated by the VLA itself and which should disappear when the EVLA is completed. Unless otherwise stated, the AIPS package (Greisen 2003) was used to reduce these data.
A faceted, low resolution image (90 ′′ clean beam) with a radius of 15 degrees was made to find interfering sources far from the area of interest. Facets centered on all very bright NVSS sources (> 30 Jy) out to 100 degrees from the field center were also included in this exploratory image. From this search 288 facets, each with 500 × 500 pixels, were chosen to cover a central region 93 ′ in radius and all the other bright sources found in the low resolution search. The facets were defined using the task SETFC which creates a set of overlapping circular regions within the square facets to cover the entire desired field. Then IMAGR was used to deconvolve all the facets together, using the standard Cotton-Schwab-Clark clean algorithm (Schwab 1984). The cell size for the final image is 2 ′′ and the clean beam size is 6.37 ′′ ×5.90 ′′ pa= 86 • .
Clean images from the first day of the observations were then used as fiducial models for each of the other days. Phase and amplitude calibrations were made of each of the other days using the clean components from the first day images. The A configuration data for each IF and polarization were then combined into a smaller, averaged dataset using STUFFR and images for the full datasets were made. The C configuration data were also calibrated using the full A configuration images. The A and C datasets were then combined using DBCON and images were made separately for each IF and polarization.
After making these images there remained some significant residual structures in the central two degrees of the image due to bright sources located outside the central region. These residuals are likely due to 1) differences in the primary beam patterns from antenna to antenna due to the very simple dipole feeds used on the VLA and 2) the rotation with parallactic angle of the sensitivit
…(Full text truncated)…
📸 Image Gallery
Reference
This content is AI-processed based on ArXiv data.
