Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.atnf.csiro.au/people/Ray.Norris/papers/n265.pdf Äàòà èçìåíåíèÿ: Wed Mar 23 01:18:47 2011 Äàòà èíäåêñèðîâàíèÿ: Fri Feb 28 08:09:12 2014 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: earth's atmosphere

Finding AGN with wide-field VLBI observations

arXiv:1011.2404v1 [astro-ph.CO] 10 Nov 2010

Enno Middelberg
Astronomisches Institut der Ruhr-UniversitÄt Bochum, UniversitÄtsstr. 150, 44801 Bochum, Germany E-mail: middelberg@astro.rub.de

Adam Deller
National Radio Astronomy Observatory, PO Box 0, Socorro, NM, 87801, USA E-mail: adeller@nrao.edu

John Morgan
Curtin University of Technology, GPO BOX U1987, Perth, WA 6845, Australia E-mail: john.morgan@curtin.edu.au

Helge Rottmann
Max-Planck-Institut fÝr Radioastronomie, Auf dem HÝgel 69, 53121 Bonn, Germany E-mail: rottmann@mpifr-bonn.mpg.de

Walter Alef
Max-Planck-Institut fÝr Radioastronomie, Auf dem HÝgel 69, 53121 Bonn, Germany E-mail: alef@mpifr-bonn.mpg.de

Steven Tingay
Curtin University of Technology, GPO BOX U1987, Perth, WA 6845, Australia E-mail: steven.tingay@icrar.org

Ray Norris
Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia E-mail: ray.norris@csiro.au

Uwe Bach
Max-Planck-Institut fÝr Radioastronomie, Auf dem HÝgel 69, 53121 Bonn, Germany E-mail: bach@mpifr-bonn.mpg.de

Walter Brisken
National Radio Astronomy Observatory, PO Box 0, Socorro, NM, 87801, USA E-mail: wbrisken@nrao.edu

Emil Lenc
Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia E-mail: emil.lenc@csiro.au

c Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence.

http://pos.sissa.it/


VLBI observations are a reliable method to identify AGN, since they require high brightness temperatures for a detection to be made. However, because of the tiny fields of view it is unpractical to carry out VLBI observations of many sources using conventional methods. We used an extension of the DiFX software correlator to image with high sensitivity 96 sources in the Chandra Deep Field South, using only 9h of observing time with the VLBA. We detected 20 sources, 8 of which had not been identified as AGN at any other wavelength, despite the comprehensive coverage of this field. The lack of X-ray counterparts to 1/3 of the VLBI-detected sources, despite the sensitivity of co-located X-ray data, demonstrates that X-ray observations cannot be solely relied upon when searching for AGN activity. Surprisingly, we find that sources classified as type 1 QSOs using X-ray data are always detected, in contrast to the 10 % radio-loud objects which are found in optically-selected QSOs. We present the continuation of this project with the goal to image 1450 sources in the Lockman Hole/XMM region.

10th European VLBI Network Symposium and EVN Users Meeting: VLBI and the new generation of radio arrays September 20-24, 2010 Manchester Uk


Speaker.

c Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence.


Finding AGN with wide-field VLBI observations

Enno Middelberg

1. Introduction
Extragalactic radio surveys are an increasingly popular method to investigate the evolution of star-forming and active galaxies. However, their resolution is of the order of arcseconds and hence not sufficient to separate the two emission processes. VLBI observations are only sensitive to brightness temperatures of around 106 K and can hence unambiguously identify AGN, provided the inferred luminosity exceeds 2 â 1021 W Hz-1 [5]. The high resolution and fringe rates of VLBI impose severe logistical constraints on wide-field imaging, however, and typical VLBI fields of view at cm wavelengths are around 10 arcseconds in diameter and hence unsuitable for surveying large numbers of objects. Here we report on the first use of a new multi-phase centre mode for the software correlator DiFX [1], which allows one to produce normal-sized VLBI data sets for a set of known target locations, and hence to image hundreds of objects with a single observing run.

2. The multi-phase centre mode of DiFX
The challenge in wide-field VLBI observations is to overcome the effects of time and bandwidth averaging, which reduce the amplitude of sources away from the phase centre. The obvious solution is to increase the temporal and spectral resolution at the correlation stage, but this results in TB-sized visibility data sets, which are difficult to process even on large computers. Whilst this approach has recently been used by Wucknitz et al. and Morgan et al. (these proceedings), we have developed an extension to the DiFX software correlator [2], which processes multiple phase centres in a single pass. In this mode, the correlation is initially performed with high frequency resolution which is sufficient to minimise bandwidth smearing. Periodically, but still frequently enough to minimise time smearing, the phase centre of the correlated data is shifted from its initial location (which is usually the pointing centre) to a target source location. This shift requires rotating the visibility phases of each baseline by an amount equal to the difference in the geometric delay between the final and initial source directions, multiplied by the sky frequency. In effect, this corrects for the "unapplied" differential fringe rotation between the final and initial source directions. This phase shift is repeated for each desired source direction. After the phase shift is applied, the visibilities are averaged in frequency and continue to be averaged in time. Eventually, this results in an array of "normal­sized" visibility datasets, with one dataset per target source. The field of view of each of these datasets is of the order of 13 , at which point bandwidth and time smearing would reduce the observed amplitudes by 5 %. These data sets can be calibrated using standard methods. Furthermore, since the phase response of a parabolic telescope is constant within the primary beam, and since the geometric delays have been taken care of at the correlation stage, the amplitude, phase, and delay corrections can simply be copied from one data set to another.

3. Observations and calibration
On 3 July 2007 we have observed with the Very Long Baseline Array (VLBA) at 1.4 GHz a single pointing centred on the original CDFS [7]. We used 8 dual-polarisation IFs with 8 MHz 2


Finding AGN with wide-field VLBI observations

Enno Middelberg

bandwidth each and two-bit sampling, resulting in a recording bitrate of 512 Mbps, which was the highest possible bitrate of the VLBA at that time. The low declination of -27 allowed us to observe the target for 9 h only, yielding a total of 178 baseline-hours from which we expect image sensitivities of 50 µ Jy (note that the low elevations caused the system temperatures to be significantly higher than at zenith during substantial parts of the observations). A six-minute phase reference cycle time was used, with 5 min allocated to the target and 1 min to the nearby strong (600 mJy) calibrator source NVSS J034838-274914. The data were recorded on disk and shipped to the Max-Planck-Institut fÝr Radioastronomie for correlation. Correlation was carried out on a 22-node computer cluster with 176 compute cores using a development version of the DiFX software correlator. From previous ATCA observations [9] the positions of 96 sources within the VLBA's primary beam were known, and so 96 data sets were produced using the new multi-phase centre mode of DiFX. Calibration followed the standard steps for a phase-referencing VLBI observation. After phase-referencing, the target S503 was found to be bright enough for self-calibration, which improved the coherence of the data. The S503 calibration tables were then copied to the other data sets and all sources were imaged. Unlike in compact-array interferometry, where the primary beam correction is carried out in the image plane, we have corrected for primary beam attenuation by calculating visibility gains. This is possible only because the fields of view are very small in our observations and so the attenuation due to the primary beam does not vary significantly across the images. We have calculated the primary beam attenuation at each station as a function of time and frequency (the primary beam size changes with frequency, and since the VLBA antennas have feeds offset from the optical axis they suffer from beam squint, which also scales with frequency).

4. Results
All sources were imaged using natural and uniform weighting (yielding restoring beams of 28.6â9.3 mas2 and 20.8â5.9 mas2 , respectively), and using a 30 % Gaussian taper at a (u, v) distance of 10 M , which resulted in a restoring beam of 55.5â22.1 mas2 . Out of the 96 targets, 20 were detected reliably, and one more target was tentatively detected. We show here contour plots of three detected targets as an illustration of the image quality. We have cross-identified the 96 targets with the deep X-ray data in the area [6, 7], to obtain as much evidence for AGN activity as possible. X-ray observations are claimed to be a very direct tracer of AGN activity [e.g., 8]. In the standard picture of AGN, the X-ray emission originates very close to the supermassive black hole, is relatively unabsorbed at energies of a few keV, and there is little contamination from other sources, such as stars. Where redshifts were available, we used the criteria by [10] to determine an object's nature using the X-ray data alone. Surprisingly, we have detected with the VLBA 7 sources which have no X-ray counterparts. One of these sources is in a region covered by a 2 Ms Chandra exposure [7]. All these targets have redshifts exceeding 0.1, hence their radio luminosities suggest that they are AGN and not starbursts [5]. We also investigated the detection rate of those VLBI-detected targets with available X-ray counterparts, as a function of X-ray source type. We found that all sources classified as type 1 3


Finding AGN with wide-field VLBI observations

Enno Middelberg
PLot file version 1 created 14-MAY-2010 14:25:06 CONT: S393 IPOL 1382.490 MHZ S393 uni pb.ICL001.1 -27 46 48.10

-27 46 48.10

PLot file version 1 created 14-MAY-2010 09:26:37 CONT: S393 IPOL 1382.490 MHZ S393 low pb.ICL001.1 -27 46 48.10

PLot file version 1 created 12-MAY-2010 14:25:28 CONT: S393 IPOL 1382.490 MHZ S393 full pb.ICL001.1

48.15

48.15

48.15

48.20 DECLINATION (J2000) DECLINATION (J2000)

48.20 DECLINATION (J2000)

48.20

48.25

48.25

48.25

48.30

48.30

48.30

48.35

48.35

48.35

48.40

48.40

48.40

48.45

48.45

48.45

48.50 03 32 01.560

01.555 01.550 01.545 01.540 RIGHT ASCENSION (J2000) Cont peak flux = 1.0872E-03 JY/BEAM Levs = 2.215E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31) version 1 created 14-MAY-2010 09:26:39 PLot file CONT: S503 IPOL 1382.490 MHZ S503 low pb.ICL001.1

01.535

03 32 01.560

01.555 01.550 01.545 01.540 RIGHT ASCENSION (J2000) Cont peak flux = 6.2810E-04 JY/BEAM Levs = 1.958E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31) version 1 created 12-MAY-2010 14:25:30 PLot file CONT: S503 IPOL 1382.490 MHZ S503 full pb.ICL001.1

01.535

03 32 01.560

01.555 01.550 01.545 01.540 RIGHT ASCENSION (J2000) Cont peak flux = -4.6887E-04 JY/BEAM Levs = 2.361E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31) version 1 created 14-MAY-2010 14:25:09 PLot file CONT: S503 IPOL 1382.490 MHZ S503 uni pb.ICL001.1

01.535

-27 48 42.05

-27 48 42.05

-27 48 42.05

42.10

42.10 DECLINATION (J2000) DECLINATION (J2000)

42.10

DECLINATION (J2000)

42.15

42.15

42.15

42.20

42.20

42.20

42.25

42.25

42.25

42.30

42.30

42.30

42.35

42.35

42.35

42.40 03 33 10.210 10.205 10.200 10.195 10.190 RIGHT ASCENSION (J2000) Cont peak flux = 1.9582E-02 JY/BEAM Levs = 2.566E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31, version 1 created 14-MAY-2010 PLot file 16, 22.63, 32, 45.25, 64, 90.51) 09:26:47 CONT: S423 IPOL 1382.490 MHZ S423 low pb.ICL001.1 10.185

42.40 03 33 10.210 10.205 10.200 10.195 10.190 RIGHT ASCENSION (J2000) Cont peak flux = 1.2857E-02 JY/BEAM Levs = 2.243E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31, version 1 created 12-MAY-2010 PLot file 16, 22.63, 32, 45.25, 64, 90.51) 14:25:41 CONT: S423 IPOL 1382.490 MHZ S423 full pb.ICL001.1 10.185

42.40 03 33 10.210 10.205 10.200 10.195 10.190 RIGHT ASCENSION (J2000) Cont peak flux = 9.2180E-03 JY/BEAM Levs = 2.496E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31, version 1 created 14-MAY-2010 PLot file 16, 22.63, 32, 45.25, 64, 90.51) 14:25:26 CONT: S423 IPOL 1382.490 MHZ S423 uni pb.ICL001.1 10.185

-27 47 18.55

-27 47 18.55

-27 47 18.60

18.60

18.60

18.65 DECLINATION (J2000) DECLINATION (J2000)

18.65 DECLINATION (J2000)

18.65

18.70

18.70

18.70

18.75

18.75

18.75

18.80

18.80

18.80

18.85

18.85

18.85

18.90

18.90

18.90

18.95 03 32 18.035

18.030 18.025 18.020 18.015 RIGHT ASCENSION (J2000) Cont peak flux = 4.7929E-04 JY/BEAM Levs = 1.885E-04 * (-1, 1, 1.414, 2, 2.828)

18.010

18.95 03 32 18.035

18.030 18.025 18.020 18.015 RIGHT ASCENSION (J2000) Cont peak flux = 3.6704E-04 JY/BEAM Levs = 1.654E-04 * (-1, 1, 1.414, 2, 2.828, 4)

18.010

18.95 03 32 18.035

18.030 18.025 18.020 18.015 RIGHT ASCENSION (J2000) Cont peak flux = -4.5892E-04 JY/BEAM Levs = 2.077E-04 * (-1, 1, 1.414, 2, 2.828)

18.010

Figure 1: Contour plots of three of the detected sources. Three images per source are shown in a row. Left column: naturally-weighted images; middle column: untapered, naturally-weighted image; right column: uniformly-weighted image. Positive contours start at three times the rms level of the images and increase by factors of 2. One negative contour is shown at three times the rms.

QSOs by the criteria by [10] were detected (there were no type 2 QSOs in our sample). This is in stark contrast to optically selected QSOs, of which typically only a fraction of 10 % is radio-loud. We investigated potential differences between VLBI-detected and undetected sources as a function of spectral index (using the data by [4]) and redshift, but found no significant differences. One detected source in our sample, S423, shows no indication of AGN activity in optical and X-ray observations. However, it is clearly detected, and given its redshift of 0.73 it must contain an AGN. Another object, S443, has been tentatively detected. However, the VLBI core is offset from the centre of the galaxy and can therefore not be taken as evidence for an AGN. We note that the radio luminosity of this source, 5 â 1021 W Hz-1 , is similar to one of the brightest radio supernovae, SN1986J. We therefore consider it possible that S443 is a radio supernova. 4


Finding AGN with wide-field VLBI observations

Enno Middelberg

PLot file version 1 created 22-OCT-2010 12:42:26 CONT: L0420 IPOL 1382.490 MHZ L0420 CL4.ICL001.1
57 09 50.38

PLot file version 1 created 22-OCT-2010 12:42:27 CONT: L0607 IPOL 1382.490 MHZ L0607 CL4.ICL001.1

57 08 54.90
50.36

54.88
50.34

54.86 DECLINATION (J2000) 54.84 54.82 54.80 54.78
50.24 DECLINATION (J2000) 50.32 50.30 50.28 50.26

54.76
50.22

54.74
50.20

54.72 10 51 22.085 22.080 22.075 22.070 22.065 RIGHT ASCENSION (J2000) Cont peak flux = 3.0718E-03 JY/BEAM Levs = 1.307E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31, 16, 22.63, 32)
10 51 52.385 52.380 52.375 52.370 RIGHT ASCENSION (J2000) Cont peak flux = 1.2183E-03 JY/BEAM Levs = 1.087E-04 * (-1, 1, 1.414, 2, 2.828, 4, 5.657, 8, 11.31) 52.365

Figure 2: Contour plots of two targets detected in the Lockman Hole East. Contours are drawn at three times the rms level, and one negative contour is shown.

5. Conclusions and outlook
This pilot project can be regarded as a success, and the multi-phase centre mode is now publicly available at the VLBA. We have therefore embarked on a project to carry wide-field VLBI even further, using mosaicing of separate pointings, to increase the sensitivity of the observations over wide areas. We have observed with the VLBA the three pointings observed by [3] with the VLA, resulting in an on-axis sensitivity of 23 uJy in each pointing. Around 330 phase centres were correlated in each run, targeting only those 508 sources with a 1.4 GHz flux density of more than 100 µ Jy. The observations resulted in a total of 1 TB of visibility data. In a follow-up project we will attempt to image all 1450 sources with a sensitivity on par with the VLA observations, when the VLBA recording rate has been increased to at least 2 Gbps. The calibration of the centre pointing has been finished, but the images have not yet been analysed. We show in Fig. 2 two example images of interestingly-looking objects. The image fidelity is higher than that achieved with the CDFS observations, something we attribute to the higher elevation, and consequently better uv-coverage and calibration transfer, of the Lockman Hole East field.

References
[1] [2] [3] [4] Deller, A. T. et al., 2007. PASP, 119, 318 Deller, A. T. et al., 2010. ApJ, submitted Ibar, E. et al., 2009. MNRAS, 397, 281 Kellermann, K. I. et al., 2008. ApJS, 179, 71 [5] Kewley, L. J. et al., 2000. ApJ, 530, 704 [6] Lehmer, B. D. et al., 2005. ApJS, 161, 21 5

[7] Luo, B. et al., 2008. ApJS, 179, 19 [8] Mushotzky, R., 2004. In A. J. Barger, ed., Supermassive Black Holes in the Distant Universe, vol. 308 of Astrophysics and Space Science Library, p. 53 [9] Norris, R. P. et al., 2006. AJ, 132, 2409 [10] Szokoly, G. P. et al., 2004. ApJS, 155, 271