University of Groningen
Nowhere to hide: identifying AGN in the faint radio sky
Radcliffe, Jack Frederick
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Radcliffe, J. F. (2019). Nowhere to hide: identifying AGN in the faint radio sky. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
6
Conclusions
In this thesis, we have explored multiple concepts regarding wide-field VLBI and the nature of the compact radio source population. While this has brought some interesting conclusions, these studies have opened up some thought-provoking questions to be investigated with further observations and analysis. In the following sections, the main conclusions of this thesis are summarised, while future prospects are discussed in the next chapter.6.1.
Chapter 2: Multi-source self-calibration
This chapter is based upon the published paper by Radcliffe et al. (2016). In this chapter, we introduced the multi-source self-calibration (MSSC) technique for use in wide-field VLBI observations. This calibration technique is designed to correct for residual phase errors that exist after standard phase referencing. In standard VLBI observations, the target field often does not have sufficient compact flux in order to perform self-calibration. However, MSSC uses multiple sources within the primary beam to provide enough compact flux in order to provide sufficient S/N for self-calibration to be successful. This algorithm utilisesuv stacking to achieve this. For this work, we used early data from the GOODS-N wide-field VLBI project (P.I. Garrett) to test and develop this calibration technique. These data were phase reference calibrated only as no primary beam correction had been developed yet.We found that The MSSC calibration technique improved the S/N of the detected sources by approximately 63% compared to standard phase referencing, and 69% compared to single-source
6
206 6.Conclusions
calibration. In addition we also provided the first publicly available implementation of this code for use in the AIPS data reduction package.
6.2.
Chapter 3: Nowhere to Hide - I. Initial catalogue
and radio properties
In this chapter, based upon the published paper by Radcliffe et al. (2018), we introduced the first full data release from the wide-field VLBI survey of the GOODS-N field. This chapter presented the first 24 hrs of observations, that targeted∼ 700separate phase centres and comprises of known VLA radio sources and sub-mm sources. As a result, a 7.05 area was contiguously covered by VLBI coverage with time and bandwidth smearing restricted to< 10%. Outside this area, only previously detected VLA sources were targeted in order to prevent unnecessary computing overheads during correlation.
As part of this chapter, we presented the first publicly available primary beam correction scheme for the European VLBI Network. This code uses the latest beam information from the EVN to provide primary beam corrections from multiple-phase centre correlated observations using the EVN.
Using the primary beam correction and MSSC calibration techniques,this chapter presents a catalogue of 31 faint VLBI-detected objects to ∼ 9 µJy beam−1 r.m.s. noise levels and radio luminosities of the order1022W Hz−1. This is almost3×the number of sources detected in previous wide-field VLBI surveys of the GOODS-N field. We find a detection fraction of25+5
−4%, which is broadly consistent with other wide-field VLBI
surveys at these flux densities.
We presented the radio properties and redshifts of these VLBI-selected objects using the vast amount of ancillary data available. We find that the VLBI-selected sources span a large range of redshifts (0.11-3.44), with brightness temperatures in the range of106-109K. In line with other VLBI surveys, we found that a large proportion (∼ 60%) of the radio emission in these objects are confined to milliarcsecond scale cores, which may hint at the existence of radio-loud, but core-dominated systems at high redshift, which may be equivalent to the population of FR0 objects in the local universe.
6
6.3.
Chapter 4: Nowhere to Hide - II. AGN selection
techniques and host galaxy properties
In this chapter, which will be submitted to Astronomy & Astrophysics, we continued our analysis of the VLBI-selected AGN population introduced in chapter 3. Using deep, multi-wavelength data in the GOODS-N field we conducted a test of the various AGN classification techniques used in other deep field studies. These included 1) searching for MIR power law AGN using theSpitzer and WISE telescopes, which search for excess MIR emission caused by the reprocessing UV photons by the AGN torus; 2) X-ray AGN using deep 2 Ms X-ray observations using theChandra telescope; and 3) radio-excess AGN, which uses the far-IR radio correlation to search for radio emission above what would be expected from star-formation alone. In addition, we tested some techniques rarely used in deep-field studies, namely radio morphologies and variability (from Chapter 5). In total, 14 different AGN classification techniques were analysed.
We reveal that no one single AGN classification technique can reliably identify all VLBI-selected sources. We find that radio variability is the least reliable, classifying around22%, which is most likely due to the sparsely sampled epochs (only 5 epochs across 22 years). This is closely followed by infra-red (20-40%), and X-rays (59%). The radio morphology classification technique works particularly well (82%), which bodes well for the upcoming e-MERGE survey that will greatly extend and improve upon this technique. We find that radio-excess measures are the most reliable (∼ 80-90%). However, the most reliable measure, which utilises the bolometric IR luminosity, requires detections in multiple bands, which limits it to just a subset of VLBI sources. After restricting the VLBI-selected sample to a region where the main IR, X-ray and radio-excess classification schemes can be evaluated for all VLBI sources, we show that the combination of these can identify all VLBI-selected sources as AGN. This independently verifies the approaches used in other deep-field studies in their quest to obtain a complete AGN sample. In particular, we find that around 20% of the VLBI sources have signatures in all three classification schemes, which alludes to the AGN dominating the bolometric luminosity of the source. These are most likely radiatively-efficient AGN with radio jets. In addition, we find around 40% are probably radiatively-inefficient AGN, hosted by large ellipticals with no evidence of a torus, as shown by their lack of IR-AGN signatures.
We find that radio-excess and VLBI-selected AGN account for around 16.7% and 5% of the total AGN content in GOODS-N. Whilst the differences between the radio-excess and VLBI observations can be explained by differing sensitivity and the resolving out of flux density, the differences between VLBI and the X-ray sources cannot. Note that the differences between infra-red AGN and VLBI are out of the scope of this
6
208 6.Conclusions
chapter. To determine whether, with deeper VLBI observations, we would detect the missing X-ray AGN, we employed stacking using the 1.5 GHz VLA radio data. Whilst around 30% of the X-ray sources have VLA counterparts (to a 5σdetection threshold of 10µJybeam−1), the remaining sources had no radio counterparts. We find that stacking upon these produces a positive signal with an integrated flux density of∼ 2 µJy. This implies that deep SKA surveys will be expected to have radio counterparts to the majority of X-ray AGN.We find that the radio flux density is uncorrelated with X-ray luminosity and this flux density is found to be on the FIR-radio correlation. This implies that the radio-emission in these X-ray AGN have a star-formation origin. Hence, further and deeper VLBI observations will not be able to categorise the majority of these sources as AGN.
6.4.
Chapter 5: The transient and variable
µ
Jy radio
sky across multiple decades
In the chapter, which has been submitted to MNRAS, we investigated the long-term variability of radio sources, at µJyflux densities, using multiple VLA epochs spread across 22 years. The existence of a variable object in the radio is another indicator of an extremely small radio emitting size, and therefore a high brightness temperature, which can only be due to AGN or supernovae. Previous variability studies are mainly probing the mJy sensitivity regime, and were often restricted to short timescales or a small number of sources on longer timescales. In this chapter, we used the deep VLA data available in the GOODS-N field to investigate the variable characteristics of around 480 sources between the years of 1996 and 2018.
The five epochs spread across this time-range were reduced and calibrated. A CASA-compatible peeling routine to remove off-axis sources was also presented. Using these epochs, we find a total of 10 unique variable sources, of which many are variable between multiple epochs. This corresponds to a detection fraction of2%of the total persistent radio source population. This is consistent with other variability surveys at higher flux densities. Therefore,we find that the variable radio sky remains constant in the µJyregime continuing the trend observed in other studies.
Using short (∼few days) spaced epochs,we reveal that the majority (70%) of the sources showing long-term variable behaviour is actually induced by rapid short-term variable behaviour, which is most likely due to scintillation effects. This shows that long term variable behaviour is rare, and must occur in≤ 1%of the persistent radio source population across decadal timescales. This could be due to the change from AGN-dominated to star-formation dominated systems at these flux density regimes, or
6
it could hint at the peak of AGN variability being in-between the short (day) timescales and long (decadal) timescales as suggested by other studies.
Using multi-wavelength diagnostics,we find evidence that the variable behaviour is AGN-related in the majority of sources and 80% of sources have VLBI counterparts. No sources show evidence of IR-AGN, which could be due to the lack of tori in these objects and could mean that the variable source population corresponds to the core-dominated FR-I sources as postulated by wide-field VLBI surveys and semi-empirical simulations. We find no firm evidence of transient behaviour. However, one source is signif-icantly offset from the optical maximum. This could be either be a radio supernova (the furthest ever discovered atz = 0.07) or an off-axis low-luminosity AGN. However, the lack of further information prevents us from reaching a final consensus upon the nature of this source.
Bibliography
Radcliffe J. F., Garrett M. A., Beswick R. J., Muxlow T. W. B., Barthel P. D., Deller A. T., Middelberg E., 2016, A&A, 587, A85
