TL;DR
This paper presents the first comprehensive catalog of extragalactic millimeter-wave point sources from the South Pole Telescope (SPT) survey, covering 771 square degrees. By analyzing sources across three frequency bands (95, 150, and 220 GHz), the authors classify them into two distinct populations – active galactic nuclei (AGN) emitting synchrotron radiation and star-forming galaxies emitting thermal dust radiation – and provide detailed number counts for each. These new measurements reveal significant discrepancies with existing galaxy population models, highlighting the need for refined theoretical frameworks to understand these cosmic sources.
What Problem Does This Paper Address?
Understanding the extragalactic sky at millimeter wavelengths is crucial for several reasons. These wavelengths are sensitive to both the synchrotron emission from the jets of active galactic nuclei (AGN) and the thermal emission from dust in star-forming galaxies. These sources, while fascinating in their own right, also act as foreground contaminants for cosmic microwave background (CMB) observations, particularly for studying the Sunyaev-Zel'dovich (SZ) effect in galaxy clusters.
Previous millimeter-wave surveys have provided catalogs and number counts, but often with limited sky coverage, fewer frequency bands, or less precise source classification. This paper aims to address these limitations by leveraging the sensitive, multi-frequency data from the SPT-SZ survey to create a large, well-characterized catalog of point sources. Specifically, the authors seek to:
- Create a robust point-source catalog: Detect and catalog sources with high significance across multiple frequency bands.
- Classify sources into distinct populations: Differentiate between synchrotron-dominated (AGN) and dust-dominated (star-forming galaxy) sources based on their spectral properties.
- Derive accurate number counts: Quantify the number of sources per unit flux density and solid angle for each population and frequency band.
- Test and constrain galaxy population models: Compare the derived number counts with theoretical models to identify areas where current models fall short and to guide future theoretical development.
Methods
The authors employ a sophisticated multi-stage approach to identify, classify, and characterize the millimeter-wave sources:
- Data Reduction and Mapmaking: Raw time-ordered data from the SPT detectors are meticulously filtered to remove atmospheric and instrumental noise. This involves techniques like Fourier-domain filtering and Legendre polynomial subtraction. The filtered data is then used to construct maps of the sky.
- Source Finding: A matched-filter technique is applied to the maps to enhance the signal-to-noise ratio of point-like sources. The CLEAN algorithm, commonly used in radio interferometry, is then employed to locate and extract sources from these filtered maps. A detection threshold of 4.5σ is used for the catalog.
- Flux Deboosting: A Bayesian deboosting procedure is crucial to correct for the "flux boosting" effect, where fainter sources are more likely to be overestimated due to positive noise fluctuations. This method simultaneously considers data from all three frequency bands, accounting for correlations between them and using prior information from number count models.
- Source Classification: Sources are classified as either synchrotron-dominated or dust-dominated based on their spectral indices derived from the deboosted fluxes. A probabilistic approach is used, with a local minimum in the posterior distribution of spectral indices serving as the classification threshold.
- Number Counts Calculation: Number counts are derived using a bootstrap method. This involves creating many mock source catalogs by resampling the deboosted flux posteriors and then calculating the counts for each mock catalog. These counts are corrected for completeness and then combined across different fields.
Datasets
- South Pole Telescope (SPT) Sunyaev-Zel'dovich (SPT-SZ) Survey Data: The primary dataset consists of observations from the SPT at 95, 150, and 220 GHz, covering a total area of 771 deg². Data from both 2008 and 2009 observing seasons are utilized.
- External Catalogs for Cross-matching: To aid in source identification and classification, the authors cross-match their detected sources with several established astronomical catalogs:
- Radio: Sydney University Molonglo Sky Survey (SUMSS), Parkes-MIT-NRAO (PMN) Southern Survey, Australia Telescope 20 GHz Survey (AT20G).
- Infrared: IRAS Faint Source Catalog (IRAS-FSC), Wide-field Infrared Survey Explorer (WISE), AKARI/FIS Bright Source Catalog, AKARI/IRC Point Source Catalog.
- Other: Planck Catalog of Compact Sources (PCCS), ROSAT All-Sky Survey (RASS).
Key Insights and Ideas
- Two Distinct Populations Emerge Clearly: The multi-frequency data from SPT allows for a clear separation of sources into two dominant populations: synchrotron emitters (AGN) and thermal dust emitters (star-forming galaxies). This classification is robust and forms the basis for much of the analysis.
- Unidentified Sources are Key: A significant fraction of sources (189 synchrotron and 189 dusty) remain unidentified after cross-matching with external catalogs. The dusty, unidentified sources are particularly interesting as strong candidates for high-redshift, strongly lensed submillimeter galaxies (SMGs).
- Models Fall Short: None of the current galaxy population models, for either synchrotron or dusty sources, provide a good fit to the measured number counts across all three frequency bands. This suggests that our understanding of these source populations and their evolution needs significant refinement.
- Deboosting is Crucial for Accurate Counts: The paper emphasizes the importance of the Bayesian deboosting procedure to correct for flux boosting, which is particularly significant at low signal-to-noise ratios. This ensures that the derived number counts are more representative of the true source population.
- High-Redshift Dusty Galaxies are Abundant: The number counts for dust-dominated sources, especially when compared to lensed models, strongly suggest the presence of a significant population of high-redshift, strongly lensed SMGs.
Key Results
- Catalog Size: A total of 1545 sources are detected above a 4.5σ significance level in at least one band.
- Source Classification: The catalog comprises 1238 synchrotron-dominated sources and 307 dust-dominated sources.
- Unidentified Sources: 189 synchrotron and 189 dusty sources have no counterparts in the external catalogs.
- Number Counts: Differential number counts are derived for each population down to flux limits of approximately 11 mJy (95 GHz), 4 mJy (150 GHz), and 11 mJy (220 GHz).
- Model Comparison: The paper presents detailed comparisons of the measured number counts with various theoretical models, highlighting significant discrepancies across all bands and for both source populations. For instance, the De Zotti et al. (2005) synchrotron model overpredicts bright blazar counts, while dust models often struggle to match the observed counts across all frequencies.
Limitations and Future Work
- Model Discrepancies: The most significant limitation highlighted is the poor fit of current galaxy population models to the SPT data. This indicates a need for improved theoretical models that better capture the physics of galaxy formation, AGN evolution, and dust emission at these frequencies.
- Statistical Fluctuations: While the survey area is large, some discrepancies between different fields or with previous results (e.g., V10) might be attributed to statistical fluctuations.
- Extended Sources: While the catalog primarily focuses on point sources, a small number of extended sources (63) were identified. Their contribution to the total counts is noted as subdominant but could be further investigated.
- Future Work: The authors emphasize that this work is a step towards analyzing the full 2500 deg² SPT-SZ survey. The final, larger catalog will provide improved statistics and tighter constraints on cosmological models and galaxy evolution. Further follow-up observations of the unidentified sources, particularly the dusty ones, are crucial for confirming their nature as lensed SMGs and for refining our understanding of high-redshift galaxy populations. The authors also plan to present a detailed comparison of source counts, power spectra, and bispectra in future work.
Summary updated