The Polarimetric Multi-Frequency Radio Sources Properties


The polarization properties of extragalactic radio sources at frequencies higher than GHz are still poorly constrained. However, their characterization would provide invaluable information about the physics of the emission processes and is crucial to estimate their contamination as foregrounds of the polarized cosmic microwave background (CMB) angular power spectrum on scales arcmin. In this contribution, after summarizing the state-of-the-art of polarimetric observations in the millimetric wavelength bands, we present our observations of a complete sample of 53 sources with mJy carried out with the Australia Telescope Compact Array between and GHz. The analysis clearly shows that polarization properties cannot be simply inferred from total intensity ones, as the spectral behaviors of the two signals are typically different.

Received Day Month Year

Revised Day Month Year

Keywords: radio galaxies; polarization; cosmic microwave background.

PACS numbers:

1 The state-of-the-art in polarimetric observations

Bright extragalactic radio source samples are mostly associated to the nuclei of active galaxies. A widely held view is that their spectra result from a superposition of a number of compact regions of different sizes, self-absorbed at different frequencies (e.g. see Ref. ? and Ref. ? for a review), associated to knot-like Doppler-boosted structures along a relativistic jet, powered by the central nuclear activity.

The characterization of the total intensity emission for large samples of radio source population is only a recent achievement, and still open to discussion. Wide-area surveys are necessary to achieve statistics on the bright less numerous samples, while high sensitivity is needed to explore the faintest samples. Only the former were available in the catalogs extracted from full-sky maps from satellite missions like WMAP (with a completeness limit of Jy at 23 GHz) and Planck (with a completeness limit of mJy at GHz ). New broad-band correlators made high sensitivity available for interferometric observations up to the mm and sub-mm regimes, allowing wide-area deep surveys to be carried out also from the ground at frequencies above GHz. Thus, the Australia Telescope 20 GHz (AT20G) Survey covered the full Southern sky with 91% completeness above mJy and completeness above mJy in regions south of declination .

By combining the ground- and satellite-based instrumental capabilities, several authors recently provided a broad-band view of the total intensity emission of the bright radio source population: it seems to be dominated by relatively young compact objects. A double power-law model is adequate to describe spectral behaviors for more than decades in frequency: this indicates that a single dominant component is responsible for the (typically optically thin) emission above GHz. Emission is typically variable: objects brighter than mJy (at GHz) on average vary their flux density (at GHz) more than over a 6 months lag, and the rate grows with frequency.

The study of the polarized emission both in frequency and space would help describing the dynamics of the jets. Synchrotron emission of each component is intrinsically highly linearly polarized (up to ) but typically observed integrated polarized fractions for compact extragalactic radio sources are rarely as high as %. They are, in fact, the result of vector averaging along the line of sight (depolarization is mostly induced by differential Faraday rotation), within the observing resolution element (in this case unresolved magnetic sub-structures cause depolarization), and within the band of the polarized components emission. Hence, polarimetric observations typically still constitute an observational challenge because of the requested high sensitivity, calibration accuracy, and the detailed knowledge of instrumental properties and systematics (see Massardi et al. in this volume). Table 1 lists some of the available multi-frequency compilations, surveys and complete samples follow-up that include polarimetric information: it does not claim to be an exhaustive picture and aims to trace the basic references in the following discussion.

For the above mentioned reasons, our current knowledge of polarimetric properties of radio sources mostly rely on GHz selected samples including the NVSS survey that covered the full sky above , remaining complete down to mJy in total intensity and with mJy in polarized emission.

References Frequency (GHz) # sources Notes Eichendorf & Reinhardt (1979) compilation of multi–frequency data Tabara & Inoue (1980) compilation of multi–frequency data Simard-Normandin et al. (1981) compilation of multi–frequency data Perley (1982) compilation of multi–frequency data Rudnick et al. (1985) compilation of multi–frequency data Aller et al. (1992) complete sample with Jy Okudaira et al. (1993) flat-spectrum sources with Jy Nartallo et al. (1998) compilation of flat-spectrum radio sources Condon et al. (1998) - NVSS complete survey down to mJy Aller et al. (1999) BLLac sources Fanti et al. (2001) CSS sample with Jy Lister (2001) complete sample with Jy Klein et al (2003) compilation of detections of the B3-VLA survey Ricci et al. (2004) complete sample with Jy Jackson et al. (2007) JVAS-CLASS surveys Massardi et al. (2008) AT20G-BSS AT20G bright sample Lopez-Caniego et al. (2009) polarization detections in WMAP maps Jackson et al. (2010) WMAP sources follow-up Murphy et al. (2010) AT20G complete survey with mJy Trippe et al. (2010) complete sample with Jy Battye et al. (2011) WMAP sources follow-up Sajina et al. (2011) AT20G sources follow-up Massardi et al. (2013) complete sample with mJy Agudo et al. (2014) complete sample of flat-spectrum sources with Jy Planck Collaboration (2015) polarization detections in Planck LFI maps (PCCS2) polarization detections in Planck HFI maps (PCCS2)

Table 1: Summary of some of the surveys in polarization available at radio frequencies (update to Table 3 in Tucci et al., 2004).

Noticeable exceptions are few high frequency-selected samples that are typically limited to flat-spectrum or bright objects.

The polarization of WMAP sources has been investigated by López-Caniego et al. (2009, Ref. ?) by using WMAP data: 14 extragalactic objects were significantly detected in polarization. Slightly larger samples were detected in the Planck maps and recorded in the “Planck Catalogue of Compact Sources” (PCCS, 2nd version), listing 122 detections down to a minimum polarized flux density of mJy at GHz but complete only to Jy. Ground-based follow-up observations of a complete sample of WMAP sources were carried out with the VLA by Jackson et al. (2010, Ref.  ?) and Battye et al. (2011, Ref. ?): polarized emission was detected for , and objects at , and GHz, respectively.

The Plateau de Bure Interferometer (PdBI) observations of Trippe et al. (2010, Ref. ?) of a  mJy complete sample of 86 sources found an average fractional polarization level of , higher for BLLac () than for QSO () or Seyfert galaxies (). The size scales relevant for the polarization emission measurements are found to be comparable to those of interest for total intensity flux density measurements.

The full AT20G catalog includes the GHz polarized intensity for sources, of which also have simultaneous polarization detections at and/or GHz, out of a total of sources. The detection limit is defined as . Sadler et al. (2006, Ref. ?) presented polarization measurements for a sample of AT20G sources brighter than mJy: () were detected at GHz, with a median fractional polarization of . Massardi et al. (2008, Ref. ?) discussed the polarization properties of the AT20G bright sample (mJy), finding polarization detections () at 20 GHz out of a total of 320 sources (), with a median fractional polarization of at 20 GHz (confirmed also by Refs. ?, ?). The spectral indices in total intensity and in polarization were found to be similar on average, but there were several objects for which the spectral shape of the polarized emission is substantially different from the spectral shape in total intensity. Several studies of radio source polarization, mostly for samples selected at 1.4 GHz and dominated by steep-spectrum objects, have reported indications that the polarization degree increases with decreasing flux density . Massardi et al. (2013, Ref. ?) analyzed high-sensitivity polarization observations (in the GHz frequency range) of a complete AT20G bright (mJy) source sub-sample and found no statistically significant relationship between the polarization fraction and the total intensity flux density, and no clear indication of trends of fractional polarization with frequency, up to GHz.

Sajina et al. (2011, Ref. ?) obtained polarization measurements with the VLA at , , , and GHz of out of the AT20G radio galaxies with mJy in an equatorial field of the Atacama Cosmology Telescope survey: polarized flux was detected at confidence level for , , , and sources, from low to high frequencies. The measured polarization fractions are typically , although in some cases they are measured to be up to . They find indications of increasing polarization fraction with frequency (confirmed also by Ref. ? in the GHz range and by Ref. ? in the GHz range). This trend is stronger for steeper spectrum sources as well as for the lower flux density sources.

Finally, it has been argued that the ordering of magnetic fields should increase in the inner regions, giving a higher polarization degree at higher frequency, while Faraday depolarization should affect more the lower frequency observations.

However, it seems clear that extrapolations from low frequencies (GHz) or from total intensity emission are inadequate to model the radio source contribution in CMB polarization maps (see Ref. ?) and that future surveys will benefit CMB polarization experiments by statistically characterizing radio source populations down to lower flux density limits. On the one hand, in fact, foreground sources are expected to be the most relevant contaminant of the CMB angular power spectrum for angular scales up to arcmin in the GHz frequency range, where Galactic foregrounds (Galactic synchrotron and thermal dust emissions) are, instead, at a minimum. On the other hand, radio mm-band facilities have reached so far sensitivities that, in principle, are useful to constraining models that predict cosmic polarization rotation (CPR), as a result of violations of fundamental principles (such as the Einstein Equivalence Principle, the Lorentz Invariance or the CPT symmetry), at the same level or even better than constraints coming from CMB observations.. Furthermore, multi-steradian samples of high frequency-selected polarized sources are also important for identifying suitable calibrators for CMB polarimetric experiments and upcoming millimeter-wave telescopes. See Massardi et al. in this volume for a short review about the role of radio sources in polarimetric cosmological investigations.

Broad frequency range and multi-epoch observations of low flux density-limited samples are now needed to complement the view of radio source population properties in polarization and provide samples useful for cosmological studies, like those that are being carried out in the framework of the Planck-ATCA Co-eval Observations project for a mJy sample selected in the AT20G catalog. The project is briefly summarized in the following section.

2 The Planck-ATCA Co-eval Observations project

The Planck-ATCA Co-eval Observations (PACO, Refs. in ?, ?, ?, ?) project, carried out with the Australia Telescope Compact Array (ATCA), observed 464 objects drawn from the Australia Telescope survey at GHz (AT20G, Refs. in ?, ?) between and GHz (below and overlapping with the two lower Planck frequency bands) in epochs during the period July 2009–August 2010, corresponding to the first two sky surveys of the satellite nominal mission. Source were scheduled for ATCA and Planck observations to occur within days of each other in order to minimize source variability effects. The sample is made of partially overlapping sub-samples, namely the bright sample (mJy), the faint sample (mJy) and the spectrally-selected sample, plus one of potentially strongly variable objects identified among the ATCA calibrators database.

The main goal of the project was to characterize the radio source population observed by the Planck satellite down to well below its detection limits (down to a factor of ) in a broad frequency range and on different epochs. Thanks to the combination of ATCA and Planck detected flux densities, we reconstructed the spectral properties of the sources in the GHz frequency range. The vast majority (%) of the sources show a remarkably smooth spectrum, well described by a double power law over the whole range. This translates in several compact objects showing a peaked spectrum with an optically thin steep behavior for frequencies GHz, probably associated to recently emitted components. This turns at odds with the criterion, well established at cm-wavelengths, according to which flat spectra are usually associated to compact sources while steep emitted synchrotron is indicative of extended sources. No further synchrotron break is observed at high frequencies. Assuming a continuous injection model, our finding can be considered an indication that the sources are young, i.e. no more than yr.

2.1 Polarimetric follow-up

In September 2014 a high sensitivity (mJy) polarimetric observation in correlator spectral setups (that allow GHz simultaneous bands centered at and GHz) was performed with ATCA (Massardi, project C2922) for a complete sub-sample of extragalactic compact radio sources of the faint (mJy) PACO sample, covering the Southern Ecliptic Pole region (ecliptic latitude ).

The choice of the region of the sky was motivated in PACO project by the Planck scanning strategy, which gives maximum sensitivity near the ecliptic poles.

The array configuration was the H214, an hybrid with a nominal spatial resolution () that spans from to arcsec for the observed frequencies.aaaIndeed, taking into account the longest baseline with the sixth antenna (Km) the spatial resolution spans from to arcsec but in this case the achievable sensitivity is reduced. All the objects in the sample are point-like.

The project was allocated in h (i.e. h per band, including overheads and calibration), during which weather conditions were very good. Every object has been observed for each spectral setup at least for min cuts, that is estimated to be enough to reach the theoretical noise level of mJy in polarized emission.

Data reduction was performed with MIRIAD software following standard polarimetric calibration procedures.bbbAccording to ATCA Users’ Guide: The outcome are high sensitivity spectral profiles in total intensity and polarization (some of them are reported in Fig. 1, together with past PACO and AT20G observations available). The detection rate in polarization is about at .

Similar to what was found in past analysis, about of the objects total intensity behavior can be fitted with a double power law. Polarimetric spectra, instead, typically show a different and sometimes more complex behavior with respect to total intensity (see Fig. 1), in agreement with Massardi et al. findings (2013, Ref. ?). As found in total intensity, a steepening of spectra is observed at higher frequencies both in total intensity and in polarization. The median fractional polarization is at GHz (in agreement with Refs. ?, ?, ?). A tiny trend with frequency is observed for the whole sample (see Fig. 2). A comprehensive analysis of the data, including spectral classifications and source counts in polarization will be presented in a future paper (Galluzzi et al., in preparation).

The project will be complemented by a second observational session scheduled with ATCA in March 2016 to observe 106 sources in the faint PACO sample with ecliptic latitude (i.e. including the already observed in the previous epoch) at all the already observed frequencies and to extend down to GHz, and by an ALMA Cycle-3 project, accepted with high-ranking, to re-observe a complete sub-sample of objects at GHz, that will provide interesting hints for cosmological applications, as described in Massardi et al. (this volume).

Fig. 1: Spectra in total intensity and polarization (error bars are not displayed since they are smaller than plotting symbols) for some sources drawn from the of the faint PACO sample, observed in September 2014. Total intensity: red crosses indicate total intensity ATCA September 2014 observations (each point represents a MHz-bin) and magenta solid lines indicate fitting curves. Median PACO flux densities (July 2009–August 2010) are indicated with orange “x”. AT20G observations (best epoch in 2004–2008) are reported with brown triangles. Polarization: black crosses and black squares indicate September 2014 observations (each point represents a GHz-bin and the GHz-full bandwidth, respectively) and blue solid lines indicate fitting curves. AT20G observations (best epoch in 2004–2008) are reported with green diamonds. Upper limits are shown as violet downwards triangles.

Fig. 2: Median fractional polarizations (error bars report inter-quartile distances) at different frequencies for our observations (black crosses) and some other samples (reported also in the table 1): the AT20G survey (blue diamonds), Klein et al. (red “x”), Agudo et al. (green triangles), and Sajina et al. (orange squares).

3 Cosmological outreach

As already stressed in the first section, extrapolations from low frequencies ( GHz) are inadequate to model the radio source contribution in CMB polarization maps and future surveys will benefit CMB polarization experiments by statistically characterizing radio source populations, as those proposed in our ALMA-Cycle 3 proposal. Moreover those stable, brighter and highly polarized sources of our sample are potentially of interest as calibrators for many ground-based facilities looking for primordial B-modes in the CMB angular power spectrum or cosmic polarization rotation, for which a clear detection has not been claimed so far and available upper limits are about . Indeed, recently the POLARBEAR Collaboration reports for the first time a sub-degree upper limit (see Ref. ?). From our data we can assess a total error on polarization angle of , by assuming that the calibration error for polarized fluxes (estimated ) is equally distributed among Stokes parameter and . However, recent measurements of polarization angles performed with VLA and ATCA are found to be in agreement within , which is a current conservative estimate of the absolute calibration error on this quantity.

4 Acknowledgments

We acknowledge financial support by the Italian Ministero dell’Istruzione, Università e Ricerca through the grant Progetti Premiali 2012–iALMA (CUP C52I13000140001). We thank the anonymous referee for the useful comments. We thank the staff at the Australia Telescope Compact Array site, Narrabri (NSW), for the valuable support they provide in running the telescope and in data reduction. The Australia Telescope Compact Array is part of the Australia Telescope which is funded by Commonwealth of Australia for operation as a National Facility managed by CSIRO.


  • 1. A. P. Marscher, Nature 288 (1980) 12
  • 2. G. De Zotti, M. Massardi, M. Negrello and J. Wall, A&A Rev. 18 (2010) 1
  • 3. R. D. Blandford and A. Königl, ApJ 232 (1979) 34
  • 4. F. Argüeso, J. González-Nuevo and L. Toffolatti, ApJ 598 (2003) 86
  • 5. G. De Zotti, R. Ricci, D. Mesa et al., A&A 431 (2005) 893
  • 6. E. L. Wright, X. Chen, N. Odegard et al., ApJS 180 (2009) 283
  • 7. B. Gold et al., ApJS 192 (2011) 15
  • 8. Planck Collaboration, arXiv:1507.02058 (2015)
  • 9. T. Murphy et al., MNRAS 402 (2010) 2403
  • 10. M. Massardi, R. D. Ekers, T. Murphy et al., MNRAS 412 (2011) 318
  • 11. M. Massardi, R. D. Ekers, T. Murphy et al., MNRAS 384 (2008) 775
  • 12. A. Sajina et al., ApJ 732 (2011) 45
  • 13. X. Chen, J. P. Rachen, M. López-Caniego et al., A&A 553 (2013) A107
  • 14. M. Massardi et al., MNRAS 455(3) (2015) 3249
  • 15. V. L. Ginzburg and S. I. Syrovatskii, ARA&A 7 (1969) 375
  • 16. B. J. Burn, MNRAS 133 (1966) 67
  • 17. W. Eichendorf and M. Reinhardt, Ap&SS 61 (1979) 153
  • 18. H. Tabara and M. Inoue, A&AS 39 (1980) 379
  • 19. M. Simard-Normandin, P. P. Kronberg and J. Neidhoefer, A&AS 43 (1981) 19
  • 20. M. Simard-Normandin, P. P. Kronberg and S. Button, ApJS 46 (1981) 239
  • 21. R. A. Perley, AJ 87 (1982) 859
  • 22. L. Rudnick, T. W. Jones, H. D. Aller et al., ApJS 57 (1985) 693
  • 23. M. F. Aller, H. D. Aller and P. A. Hughes, ApJ 399 (1992) 16
  • 24. A. Okudaira, H. Tabara, T. Kato and M. Inoue, PASJ 45 (1993) 153
  • 25. R. Nartallo et al., MNRAS 297 (1998) 667
  • 26. J. J. Condon, W. D. Cotton, E. W. Greisen et al., AJ 115 (1998) 1693
  • 27. M. F. Aller, H. D. Aller and P. A. Hughes, BAAS 31 (1999) 1397
  • 28. C. Fanti, F. Pozzi, D. Dallacasa et al., A&A 369 (2001) 380
  • 29. M. L. Lister, S. J. Tingay and R. A. Preston, ApJ 554 (2001) 964
  • 30. U. Klein, K.-H. Mack, L. Gregorini and M. Vigotti, A&A 406 (2003) 579
  • 31. R. Ricci, I. Prandoni, C. Gruppioni et al., A&A 415 (2004) 549
  • 32. N. Jackson, R. A. Battye, I. W. A. Browne et al., MNRAS 376 (2007) 371
  • 33. M. López-Caniego, M. Massardi, J. González-Nuevo et al., ApJ 705 (2009) 868
  • 34. N. Jackson et al., MNRAS 401 (2010) 1388
  • 35. S. Trippe, R. Neri, M. Krips et al. A&A 515 (2010) A40
  • 36. R. A. Battye et al., MNRAS 413 (2011) 132
  • 37. M. Massardi, S. G. Burke-Spolaor, T. Murphy et al., MNRAS 436 (2013) 2915
  • 38. I. Agudo, C. Thum, J. L. Gómez and H. Wiesemeyer, A&A 566 (2014) A59
  • 39. E. M. Sadler et al., MNRAS 371 (2006) 898
  • 40. D. Mesa, C. Baccigalupi, G. De Zotti et al., A&A 396 (2002) 463
  • 41. M. Tucci et al., MNRAS 349 (2004) 1267
  • 42. A. R. Taylor, J. M. Stil, J. K. Grant et al., ApJ 666 (2007) 201
  • 43. J. K. Grant, A. R. Taylor, J. M. Stil et al. ApJ 714 (2010) 1689
  • 44. R. Subrahmanyan et al., MNRAS 402 (2010) 2792
  • 45. I. Agudo, C. Thum, H. Wiesemeyer and T. P. Krichbaum, ApJS 189 (2010) 1
  • 46. QUIET Collaboration, arXiv:1412.1111 (2014)
  • 47. W-T Ni, Prog. Theor. Phys. Suppl. 172 (2008) 49
  • 48. V. A. Kostelecký and M. Mewes, Phys. Rev. D 80 (2009) 015020
  • 49. S. di Serego Alighieri, Int. J. Mod. Phys. D 24 (2015) 1530016
  • 50. POLARBEAR Collaboration, arXiv:1509.02461 (2015)
  • 51. M. Massardi et al., MNRAS 415 (2011) 1597
  • 52. L. Bonavera et al., MNRAS 416 (2011) 559
  • 53. A. Bonaldi et al., MNRAS 428 (2013) 1845
  • 54. N. S. Kardashev, Soviet Astronomy Vol. 6 (1962) 317
  • 55. M. Murgia, Publ. of the Astr. Soc. of Australia Volume 20, Issue 1 (2002) 19
  • 56. M. Tucci and L. Toffolatti, AdAst 2012 (2012) 624987
  • 57. D. Hanson et al., PRL 111 (2013) 141301
  • 58. PRISM Collaboration, arXiv:1306.2259 (2013)
  • 59. B. Partridge, M. López-Caniego, R. A. Perley et al., arXiv:1506.02892 (2015)
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description