Deep wide-field GMRT surveys at 610 MHz

The GMRT has been used to make deep, wide-field surveys of several fields at 610 MHz, with a resolution of about 5 arcsec. These include the Spitzer Extragalactic First Look Survey field, where 4 square degrees were observed with a r.m.s. sensitivity of about 30 microJy/beam, and several SWIRE fields (namely the Lockman Hole, ELAIS-N1 and N2 fields) covering more than 20 square degrees with a sensitivity of about 80 microJy beam or better. The analysis of these observations, and some of the science results are described.

💡 Research Summary

The paper presents the results of deep, wide‑field radio surveys carried out with the Giant Metrewave Radio Telescope (GMRT) at 610 MHz, achieving an angular resolution of about 5 arcsec. The authors targeted the Spitzer Extragalactic First Look Survey (xFLS) field and several SWIRE fields (Lockman Hole, ELAIS‑N1, and ELAIS‑N2). The xFLS field covers roughly 4 deg² and reaches an rms noise of ≈30 µJy beam⁻¹, while the SWIRE fields together span more than 20 deg² with rms sensitivities of ≈80 µJy beam⁻¹ or better.

Observations were performed using two 16‑MHz sidebands (right‑ and left‑hand circular polarization), each split into 128 spectral channels to allow precise identification and excision of narrow‑band radio frequency interference (RFI). Standard primary calibrators (3C 286, 3C 48) were observed at the start and end of each run to set the absolute flux density scale and to derive antenna‑based band‑pass corrections. A pseudo‑continuum (“channel 0”) was created by averaging central channels, and nearby secondary calibrators were observed roughly every 30 minutes to track time‑dependent phase and amplitude variations. After applying band‑pass and gain calibrations, the data were averaged down to ten channels to reduce dataset size before imaging.

Because the GMRT primary beam at 610 MHz is ≈43 arcmin (≈0.4 deg²), each field required a mosaic of many pointings. Typically 19 facets (a central facet, a surrounding ring of six, and a larger ring of twelve on a hexagonal grid) were imaged and later combined to produce a seamless map of each region. The imaging pipeline employed several cycles of self‑calibration: three phase‑only rounds with solution intervals of 10, 3, and 1 minutes, followed by a final combined amplitude‑and‑phase calibration with a 10‑minute interval. This procedure yielded dynamic ranges of several thousand to one, limited only by thermal noise away from bright sources. Near bright sources, image quality was degraded by antenna pointing offsets, especially at low elevations. The authors identified two systematic issues: (1) a ≈7 s error in the uv‑data time stamps, causing a slight rotation of the synthesized images, and (2) flux‑density discrepancies between overlapping pointings, attributed to pointing offsets. They corrected the time‑stamp error with a custom AIPS task and empirically adjusted the effective primary‑beam position to mitigate the pointing problem.

The final mosaics achieve rms noise levels of ≈30 µJy beam⁻¹ for the xFLS field (comparable to the deep VLA 1.4‑GHz survey of the same region, which required ≈200 h of observing time to reach ≈23 µJy beam⁻¹) and ≈80 µJy beam⁻¹ for the SWIRE fields (requiring ≈40 h per field). These data provide catalogues of several thousand radio sources, enabling a range of astrophysical investigations.

A key scientific result is the construction of a sample of 235 sources detected at both 610 MHz and 1.4 GHz, and also at 24 µm and 70 µm in the infrared, with spectroscopic redshifts up to z ≈ 1. By applying k‑corrections based on measured radio spectral indices and infrared spectral‑energy‑distribution models, the authors examined the infrared‑radio correlation (IRRC) as a function of redshift. No significant evolution of the IRRC is found, implying that the magnetic field strength in star‑forming galaxies has not changed appreciably since z ≈ 1.

The authors also employed a median stacking technique to probe populations below the individual detection threshold. Stacking 591 xFLS sources with 24 µm flux densities between 150 and 200 µJy yields a median 610 MHz flux density of 21 ± 2 µJy, confirming that the IRRC holds for fainter infrared sources as well.

An intriguing subset of “infrared‑faint radio sources” (IFRS) is identified: 14 objects in the xFLS field lack any Spitzer detection despite being relatively bright at 610 MHz. Optical imaging shows that eight of them are near the R‑band limit (R ≈ 24.5). Their radio morphology and lack of infrared emission suggest they are compact (≲20 kpc) Fanaroff‑Riley II radio galaxies at high redshift (z ≈ 4).

Finally, differential source counts at 610 MHz are derived over the flux range 0.27 mJy to 200 mJy. A clear flattening below ≈2 mJy is observed, consistent with a three‑component population model comprising steep‑spectrum AGN, flat‑spectrum AGN, and star‑burst galaxies undergoing pure luminosity evolution. This model simultaneously reproduces the 610 MHz counts and the well‑studied 1.4 GHz counts from the literature.

In summary, the GMRT 610 MHz surveys provide high‑resolution, deep radio imaging over large sky areas, complementing existing infrared and higher‑frequency radio data. The work demonstrates the utility of low‑frequency, wide‑field observations for studying galaxy evolution, magnetic field history, and the nature of faint radio populations, and it establishes robust data‑reduction techniques that address instrumental systematics unique to the GMRT.