[Paper Review] The 9C Survey: a deeper source count at 15 GHz

Elizabeth Waldram, G. G. Pooley|arXiv (Cornell University)|Jul 20, 2004

Radio Astronomy Observations and Technology3 citations

TL;DR

This paper presents a deeper 15 GHz radio survey using the Ryle Telescope, achieving a 5 mJy completeness limit across 12 deg², significantly improving source count statistics. It confirms a power-law source count with index −2.15 down to 5 mJy, with no evidence of turnover, and reveals increasing spectral index inversion with decreasing flux density, suggesting evolving source populations at low flux levels.

ABSTRACT

Foreground radio sources are a major contaminant for centimetre-wave cosmic microwave background (CMB) measurements and the 9C survey was set up as part of the observing strategy of the CMB telescope, the Very Small Array. Prior to this survey with the Ryle Telescope at 15 GHz there was no comparable high-frequency radio survey of any extent. Our first published source count reached a limit of ~25mJy but we have now surveyed some areas more deeply, to a completenesss limit of ~5mJy. We present the results from this deeper survey.

Motivation & Objective

• To extend the 9C survey to deeper flux limits at 15 GHz to better model foreground contamination in CMB observations.
• To improve source count statistics at low flux densities (5–25 mJy) for accurate CMB foreground subtraction.
• To investigate spectral index distributions across flux density bins to understand source population evolution.
• To support the Very Small Array (VSA) CMB observations by providing high-cadence source monitoring data.

Proposed method

• Conducted deep 15.2 GHz observations using the Ryle Telescope with a raster scanning technique over 12 deg² to achieve 5 mJy completeness.
• Performed point-source detection and flux density measurement via CLEANed maps from 72 individual pointings, with integration over multiple observation sets.
• Correlated 15 GHz sources with 1.4 GHz surveys (NVSS/FIRST) to determine spectral indices using S ∝ ν^−α.
• Used a differential source count model n(S) ≈ 51 (S/Jy)^−2.15 Jy⁻¹ sr⁻¹ to fit the data and compare with theoretical predictions.
• Identified and followed up peak sources with pointed observations to ensure reliable flux density measurements.
• Analyzed spectral index distributions across three flux density bins: 5–25 mJy, 25–100 mJy, and >100 mJy.

Experimental results

Research questions

• RQ1 Does the differential source count at 15 GHz show a turnover at flux levels below 25 mJy, as predicted by some models?
• RQ2 How do spectral index distributions vary across different flux density ranges, particularly in the 5–25 mJy range?
• RQ3 To what extent do sources with inverted spectra (α < 0) dominate at low flux levels, and what does this imply for source evolution?
• RQ4 Can the 9C survey’s deeper data constrain the Toffolatti et al. model prediction for the 15 GHz source count?
• RQ5 What is the reliability of source flux density measurements at the 5 mJy level using the rastering and CLEANing technique?

Key findings

• The differential source count at 15 GHz follows a power-law form n(S) ≈ 51 (S/Jy)^−2.15 Jy⁻¹ sr⁻¹ down to 5 mJy, with no evidence of a turnover.
• The deeper survey data lie on the same power-law curve as the original 25 mJy survey, confirming consistency across flux levels.
• 10% of sources in the 5–25 mJy range have inverted spectra (α₁.₄¹⁵.² < 0), rising to 20% in the 25–100 mJy range and 33% above 100 mJy.
• Only one source (5.3 mJy) at 15 GHz lacks a counterpart at 1.4 GHz in NVSS or FIRST, indicating high completeness in cross-identification.
• The noise level in the final raster maps is below 1 mJy, enabling reliable detection and flux measurement at the 5 mJy completeness limit.
• The spectral index distribution shows a systematic increase in the fraction of inverted-spectrum sources with decreasing flux density, suggesting a population evolution effect.