Nuclear radio activity in PDS 456 1

The nearby extreme accretion and feedback system

PDS 456: finding a complex radio-emitting nucleus

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

Jun Yang,1⋆ Zsolt Paragi,2 Emanuele Nardini,3,4 Willem A. Baan,5 Lulu Fan,6,7,8
Prashanth
1
Mohan,9 Eskil Varenius1,10 and Tao An9
Department of Space, Earth and Environment, Chalmers University of Technology, Onsala Space Observatory, SE-439 92 Onsala, Sweden
2 Joint Institute for VLBI ERIC (JIVE), Postbus 2, NL-7990 AA Dwingeloo, the Netherlands
3 Dipartimento di Fisica e Astronomia, Università di Firenze, via G. Sansone 1, I-50019 Sesto Fiorentino, Firenze, Italy
4 INAF âĂŞ Osservatorio Astrofisico di Arcetri, Largo Enrico Fermi 5, 50125 Firenze, Italy

PT
5 Netherlands Institute for Radio Astronomy ASTRON, NL-7991 PD Dwingeloo, the Netherlands
6 CAS Key Laboratory for Research in Galaxies and Cosmology, Department of Astronomy, University of Science and Technology of China,

230026 Hefei, China

7 School of Astronomy and Space Sciences, University of Science and Technology of China, 230026 Hefei Anhui, China

RI
8 Shandong Provincial Key Lab of Optical Astronomy and Solar-Terrestrial Environment, Institute of Space Science, Shandong University,

264209 Weihai, China

9 Shanghai Astronomical Observatory, Key Laboratory of Radio Astronomy, Chinese Academy of Sciences, 200030 Shanghai, China

SC
10 Jodrell Bank Centre for Astrophysics, The University of Manchester, Oxford Rd, Manchester M13 9PL, UK

U
Accepted 2020 XXX. Received 2020 YYY; in original form 2020 ZZZ

N
ABSTRACT
A
When a black hole accretes close to the Eddington limit, the astrophysical jet is often accompanied
M
by radiatively driven, wide-aperture and mildly relativistic winds. Powerful winds can produce signif-
icant non-thermal radio emission via shocks. Among the nearby critical accretion quasars, PDS 456
has a very massive black hole (about one billion solar masses), shows a significant star-forming activ-
ity (about seventy solar masses per year) and hosts exceptionally energetic X-ray winds (power up
ED

to twenty per cent of the Eddington luminosity). To probe the radio activity in this extreme accre-
tion and feedback system, we performed very-long-baseline interferometric (VLBI) observations of
PDS 456 at 1.66 GHz with the European VLBI Network (EVN) and the enhanced Multi-Element Re-
IT

motely Linked Interferometry Network (e-MERLIN). We find a rarely-seen complex radio-emitting

nucleus consisting of a collimated jet and an extended non-thermal radio emission region. The diffuse
emission region has a size of about 360 pc and a radio luminosity about three times higher than
ED

the nearby extreme starburst galaxy Arp 220. The powerful nuclear radio activity could result from
either a relic jet with a peculiar geometry (nearly along the line of sight) or more likely from diffuse
shocks formed naturally by the existing high-speed winds impacting on high-density star-forming
N

regions.
Key words: galaxies: active – galaxies: jets – galaxies: nuclei: – quasars: individual:
U

PDS 456 – radio continuum: galaxies

L
A

1 INTRODUCTION Panessa et al. 2019). When the SMBHs have accretion

rates approaching the Eddington limit, they may pro-
IN

Accreting supermassive black holes (SMBHs) can provide

duce not only jets (e.g. Giroletti et al. 2017; Yang et al.
mechanical feedback on their host galaxies via launching jets
2019, 2020a) but also extremely powerful winds (e.g.
and winds. Jets are collimated relativistic outflows emitting
Nardini et al. 2015; Tombesi et al. 2015; Fiore et al. 2017).
G

synchrotron emission (e.g. Blandford, Meier & Readhead

These winds can sweep out the interstellar medium
2019). Winds are non-jetted, relatively wider angled out-
< 0.3 c (e.g. Tombesi (e.g. Zakamska & Greene 2014), produce radio-emitting
RI

flows with a relatively low speed of ∼

shocks (e.g. Nims, Quataert & Faucher-Giguère 2015), con-
2016) and are mainly traced by optical and X-ray spec-
tribute the extragalactic gamma-ray background (e.g.
troscopic as well as radio continuum observations (e.g.
Lamastra et al. 2017) and quench star formation (e.g.
O

Kormendy & Ho 2013; Fiore et al. 2017). Observing this

complex nuclear radio activity with the very-long-baseline
⋆ E-mail: [email protected]
© The Author(s) 2020. Published by Oxford University Press on behalf of The Royal Astronomical Society. This is an Open Access article
distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
2 J. Yang et al.
interferometric (VLBI) imaging technique can provide in-
formation to constrain physical properties in the extreme (a) J1724−1443 (b) Residual map

accreting environment and help understand feedback with

the host galaxy.
The quasar PDS 456 (IRAS 17254−1413) at a red-
shift z = 0.184 is the most luminous quasar in the lo-
cal (z < 0.3) Universe (Torres et al. 1997; Simpson et al.
1999). The quasar has a bolometric luminosity Lbol ∼

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

1047 erg s−1 (Reeves et al. 2000) comparable to its Edding-
ton luminosity LEdd ∼ 1.3 × 1047 erg s−1 (Nardini et al.
2015) and a disk geometry close to face-on (Yun et al. 2004;
Bischetti et al. 2019). It is also an ultraluminous infrared
(IR) galaxy (ULIRG) with a luminosity of LFIR = 1.3 ×
1012 L⊙ (Yun et al. 2004). Compared to the known nearby

PT
critical accretion systems, its central SMBH has the high-
est mass, Mbh = 109.2±0.2 M⊙ (supplementary materials,
Nardini et al. 2015).
As a critical accretion system, PDS 456 is of great in-

RI
terest for probing powerful multi-phase winds and outflows.
< 0.3 c) X-ray
It hosts a quasi-spherical mildly relativistic ( ∼

SC
wind with a very high kinetic energy ∼0.2LEdd (Nardini et al.
2015). The power of X-ray winds positively correlates with
the X-ray luminosity, and thus these winds are likely ra-

U
diation pressure driven (Matzeu et al. 2017). From soft
X-ray (Boissay-Malaquin et al. 2019; Reeves et al. 2020)

N
to ultra-violet (O’Brien et al. 2005; Hamann et al. 2018)
bands, there are also more reports of highly blue-shifted
absorption lines at velocities consistent with the hard X-
A
ray winds. Moreover, the CO (3–2) emission line observa- Figure 1. The point-source model fitting results of the visibility
tions at ∼1 mm show that there exists not only some spa-
M
data of the calibrator J1724−1443. (a) The high-dynamical range
tially extended molecular outflows up to ∼5 kpc but also map of its complex jet structure. The full width at half maximum
high-velocity (∼800 km s−1 ) compact outflows in the nucleus (FWHM) of the synthesised beam is 29.6 mas × 6.6 mas at 73.9◦ .
(Bischetti et al. 2019). The contours are 0.05 × (−1, 1, 2, ..., 64) mJy beam−1 . The image
ED

Based on the multi-wavelength spectral energy distri- peak is 229 mJy beam−1 . With respect to the image noise level
bution (SED), PDS 456 is a radio-quiet analogue of the 0.017 mJy beam−1 , the dynamic range reaches 13 740. The inset
plots the non-optimal (u, v) coverage. (b) The noise distribution
well-known radio-loud blazar 3C 273 (Yun et al. 2004). The
in the residual map. The display range is between −0.083 to +0.070
IT

radio counterpart of PDS 456 is a steep-spectrum source mJy beam−1 .

(Yun et al. 2004) with a flux density of 25 ± 4 mJy at 1 GHz
and a spectral index of α = −0.84 ± 0.11 (Yang et al. 2019).
ED

Besides the intense active galactic nucleus (AGN), its SED

Table 1. Summary of the VLBI observations of PDS 456 at 1.66
in the far IR shows evidence for significant star-forming ac- GHz.
tivity in its host galaxy (Yun et al. 2004). The radio emission
is likely dominated by the AGN activity (Yun et al. 2004).
N

Observing date and time (UT) Participating stations

The pilot high-resolution VLBI observations at 5 GHz found
a faint jet structure consisting of a few components on the 2018 Mar 28, 02h30m–08h30m JbWbEfMcOnTrHh
U

deca-pc scale in the nuclear region, while failed to image 2020 Jan 22, 06h30m–13h00m JbWbEfOnTrHhIrSrKnDaPiCmDe
some very diffuse radio structure (∼70 per cent of its total
flux density, Yang et al. 2019). To get a complete view on
L

the potential large scale radio activity which may owe to H0 = 71 km s−1 Mpc−1 , Ωm = 0.27, ΩΛ = 0.73 is adopted;
A

an episodic jet, and understand nuclear starbursts and/or the images then have a scale of 3.1 pc mas−1 .
wind shocks, we performed new VLBI imaging observations
IN

at a suitable frequency of 1.66 GHz with the European VLBI

Network (EVN) plus the enhanced Multi-Element Remotely
2 OBSERVATIONS AND DATA REDUCTION
Linked Interferometry Network (e-MERLIN).
G

This paper is organised as follows. We describe the Table 1 gives a summary of our VLBI observations at
VLBI observations and the data reduction in Section 2 and 1.66 GHz. At first, we observed PDS 456 with the EVN
RI

present the imaging results of PDS 456 in Section 3. We on 2018 March 28. Owing to a severe instrumental problem
discuss star-formation rate based on the IR SED, physical with the digital filters, the target was re-observed with the
characteristics of the observed complex nuclear radio activ- EVN plus the e-MERLIN on 2020 January 22. In the two ex-
O

ity and potential identification with the jet, starburst and periments, the participating stations were Jodrell Bank MK
shock activity, and the accretion rate in Section 4. Through- II (Jb), Westerbork single antenna (Wb), Effelsberg (Ef),
out the paper, a standard ΛCDM cosmological model with Medicina (Mc), Onsala (On), Toruń (Tr), Hartebeesthoek
 Nuclear radio activity in PDS 456 3
(Hh), Irebene (Ir), Sardinia (Sr), Knockin (Kn), Darnhall group of delta functions, i.e., point source models, and the
(Da), Pickmere (Pi), Cambridge (Cm), Defford (De). self-calibration in difmap (Shepherd, Pearson & Taylor
Both observations were carried out in the e-VLBI mode. 1994). This is similar to the non-negative least-squares
The data transfer speeds were 1024 Mbps (16 sub-bands in algorithm (NNLS, Briggs 1995) in the image plane and
dual polarisation, 16 MHz per sub-band, 2-bit quantisation) thus can also allow users to achieve a high-dynamic-range
at the EVN stations and 512 Mbps (2 sub-bands in dual map (Yang et al. 2020b). We re-ran the fringe-fitting and
polarisation, 64 MHz per sub-band, 2-bit quantisation) at the amplitude and phase self-calibration in aips with the
the e-MERLIN stations (Kn, Da, Pi, Cm and De). The data input source model made in difmap. All these solutions

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

correlation was done in real time by the EVN software cor- were also transferred to the data of PDS 456 by the linear
relator (SFXC, Keimpema et al. 2015) at JIVE (Joint Insti- interpolation.
tute for VLBI, ERIC) using an integration time of 1 s and The total intensity and residual maps from the direct
a frequency resolution of 1 MHz. model fitting are shown in Fig. 1. The calibrator J1724−1443
The observations of the faint source PDS 456 ap- has a single-sided core-jet structure with a total flux den-
plied the phase-referencing observing technique. The pc- sity of 0.27 ± 0.02 Jy. Its radio core has a flux density of

PT
scale compact source J1724−1443 (Petrov et al. 2006), 0.23 ± 0.01 Jy and is used as the reference point in the
about 59 arcmin apart from our target, was used as the phase-referencing calibration. There were 79 point source
phase-referencing calibrator. Its J2000 position is RA = models used in the Stokes I map. Both the Stoke I and zero-
17h 24m 46.s 96654 (σra = 0.1 mas), Dec. = −14◦ 43′ 59.′′ 7609 flux density V maps have almost the same off-source noise

RI
(σdec = 0.2 mas) in the source catalogue 2016a from the level, σrms = 0.017 mJy beam−1 . The noise distribution in
Goddard Space Flight Centre (GSFC) VLBI group. The cy- the residual map of Stokes I is quite uniform in particular in

SC
cle time of the periodic nodding observations was about four the on-source region, although there is only one station (Hh)
minutes (∼0.5 min for J1724−1443, ∼2.5 min for PDS 456, on the long baselines. Because the method tried to use the
∼1.0 min for two gaps). A very bright flat-spectrum radio minimum number of point sources and the (u, v) coverage is
quasar NRAO 530 (J1733−1304, e.g. An et al. 2013) was also poor, the faint jet structure does not looks very smooth. The

U
observed to determine the instrumental phases and bandpass e-MERLIN data were excluded in the final image because
shapes. they gave some low-level noise peaks (∼0.15 mJy beam−1 ,

N
The data were calibrated using the National Radio ∼0.07 per cent of the peak brightness) in the residual map.
Astronomy Observatory (NRAO) software package Astro- We imaged the faint target PDS 456 without any
A
nomical Image Processing System (aips, Greisen 2003). self-calibration in difmap. First, we made a low-resolution
First, we reviewed the data carefully and flagged out off- clean map with the e-MERLIN and Jb. The target PDS 456
M
source data and some very low-sensitivity data. Second, we has a peak brightness of 7.6±0.2 mJy beam−1 using natural
ran a-priori amplitude calibration with properly smoothed weighting (beam FWHM: 506 × 111 mas2 in position angle
antenna monitoring data (system temperatures and gain 14◦ ), and indicates a slightly resolved structure with a to-
ED

curves) or nominal system equivalent flux densities when tal flux density of 12.3±0.3 mJy and a de-convolved size of
these monitoring data were not available. Third, we updated ∼123 mas. After the deconvolution of the major component
the position of the antenna Sr with the aips task clcor with a proper window, there is a hint of two-sided jet-like
(options ANAX and ANTP) according to the first geode- structure at position angle (PA) of about −25◦ and +145◦ in
IT

tic VLBI measurements (project code EL054, reported by the residual map but they are not bright enough (<5σ) for a
Leonid Petrov on the web1 ): J2000 epoch, position: (X, Y, confirmation as real features. This is followed up by making
ED

Z) = (+4865183.542, +791922.255, +4035136.024) m, veloc- a high-resolution map with the EVN plus e-MERLIN data.
ity: (Vx , Vy , Vz ) = (−12.15, +19.29, +10.86) mm yr−1 and axis We used σ −2 (σ, data error) as the visibility data weight and
offset 0.031 m. The corrections are (∆X, ∆Y, ∆Z) = (+0.6116, the purely natural grid weighting to clean the diffuse emis-
−0.1553, −1.0446) m for the antenna position and +0.031 m sion (peak brightness: 0.23 mJy beam−1 ) in the nuclear re-
N

for the antenna axis offset. Fourth, the general bootstrap gion. Owing to residual phase errors of the phase-referencing
phase calibrations were performed. We corrected the iono- calibration, our final image has a noise level about a factor of
U

spheric dispersive delays according to the maps of total elec- two higher than the statistical value estimated in the zero-
tron content provided by Global Positioning System (GPS) flux Stokes V map.
satellite observations, removed phase errors due to the an-
L

tenna parallactic angle variations, aligned the phases across

A

the subbands via running fringe-fitting with a two-minute

scan of the NRAO 530 data. After these calibrations, we 3 EVN PLUS E-MERLIN IMAGING RESULTS
IN

combined all the subbands, ran the fringe-fitting and ap- The total intensity maps of PDS 456 observed with the EVN
plied the solutions to PDS 456 by interpolation. Finally, plus the e-MERLIN on 2020 January 22 are shown in Fig. 2.
the bandpass calibration was performed. All the above steps The beam synthesised with natural weighting has an FWHM
G

were scripted in the ParselTongue interface (Kettenis et al. of 23 × 5.2 mas2 in PA = +73◦ . To avoid resolving out diffuse
2006). radio features, we artificially increased the beam area and
RI

We first imaged the phase-referencing source used a circular beam with a FWHM = 23 mas in the map.
J1724−1443. The imaging procedure was performed The nucleus of PDS 456 consists of two relatively compact
through a number of iterations of model fitting with a knots, denoted as the components C and NW, and a diffuse
O

emission region extending mainly toward West and South.

According to the position of its centroid, we labelled the
1 http://astrogeo.org/petrov/discussion/el054/ complex structure as the component W. The unshown image
4 J. Yang et al.

0.0 mJy/beam 1.0 The very extended component W has a brightness temper-
ature of ∼3 × 105 K.
-14 15 55.6
PDS 456
4 DISCUSSION
NW 4.1 Star formation rate based on IR photometries

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

55.8
C To estimate the star formation rate (SFR) of PDS 456, we
constructed the IR SED based on the multi-band photome-
Dec. (J2000)

tries in the literature. The photometries of the Wide-field

Infrared Survey Explorer (WISE), the Infrared Astronomi-
cal Satellite (IRAS), the Infrared Space Observatory (ISO)
and the James Clerk Maxwell Telescope (JCMT) Submil-
W limetre Common-User Bolometer Array 2 (SCUBA-2) have

PT
56.0
been retrieved from NASA/IPAC Extragalactic Database.
The Herschel photometries have been obtained by using the
Herschel Interactive Processing Environment (hipe v15.0.1).

RI
200 pc Compared to the previous SFR estimate based on SED fit-
ting by Yun et al. (2004), we add two WISE bands at 12
and 22 µm, two Herschel PACS (Photodetector Array Cam-

SC
17 28 19.805 19.795 19.785 19.775
RA (J2000) era and Spectrometer) bands at 70 and 160 µm, and three
SPIRE (Spectral and Photometric Imaging REceiver) bands
at 250, 350 and 500 µm. The inclusion of the Herschel PACS

U
Figure 2. The total intensity image of the quasar PDS 456 and SPIRE photometry is crucial for the SFR estimate as
observed by the EVN plus e-MERLIN at 1.66 GHz. A circular they map the cold dust emission peaking at 100–200 µm,

N
restoring beam with FWHM = 23 mas is used. The contours
which is directly involved in the star formation activity.
start from 0.1 mJy beam−1 (3σ) and increase by a factor of −1,
1, 2, 4, 8 and 16.
We use the Bayesian SED modelling and interpreting code
A
BAYESED (Han & Han 2012, 2014, 2019) to decompose
the IR SED by using an AGN torus model and a simple mod-
M
from the first epoch observation has a significantly lower ified blackbody model to represent the contribution from
quality owing to the limited telescopes and sensitivity, yet cold dust emission heated by a young stellar population (for
it independently confirms the existence of the components more details, see Fan et al. 2016, 2017, 2019).
ED

C, NW and W. The relatively bright spot at ∼12σ in the The derived IR luminosity of cold dust is
South, Fig. 2 was only marginally (∼3σ) detected in the ∼ 6.9 × 1011 L⊙ , which is lower than that of Yun et al.
early epoch. (2004) by a factor of about two. The difference is mainly
The circular Gaussian model fitting results in difmap due to the following two factors. First, our SED included
IT

(version 2.5e) are given in Table 2. All the errors are for- the more data points at far IR wavelengths than that of
mal uncertainties measured at the reduced χ2 = 1. Em- Yun et al. (2004). Second, we excluded the contribution of
ED

pirically, the flux density measurements have a systematic the AGN torus emission in the IR luminosity estimate. We
error of around ten per cent. Compared to the flux den- convert the derived IR luminosity of cold dust into a SFR
sity (16 ± 2 mJy at 1.66 GHz) predicted by its radio spec- of about 69 M⊙ yr−1 , assuming a Chabrier initial mass
trum (Yang et al. 2019), the VLBI map only restores ∼70 function (Chabrier 2003). We note that this SFR estimate
N

per cent. With respect to J1724−1443, the differential as- should be taken as an upper limit due to source blending in
trometry gives a J2000 position of RA = 17h 28m 19.s 78958 and the IR bands. The SFR is also in agreement with the results
U

Dec. = −14◦ 15′ 55.′′ 8520 for the peak component C. This is in reported by Bischetti et al. (2019), suggesting significant
agreement with the average position of components C1 and but no extreme star forming activity in the host galaxy.
C2 at 5 GHz reported by (Yang et al. 2019) if we consider
L

significant systematic error due to the extended source struc-

4.2 Complex nuclear radio activity
A

ture and the different (u, v) coverage, ∼0.1θ size (∼1 mas).
With respect to the component C, we reported the relative The sub-kpc scale radio nucleus of PDS 456 shows a
IN

offsets of components NW and W in Table 2. complex radio morphology. This is not unusual in ultra-
In the last two columns of Table 2, we also report radio violet to sub-mm luminous but radio weak/quiet sources
luminosity LR = νLν (ν, observing frequency) and brightness because their radio emission originate from multiple
temperature Tb , estimated (e.g., Condon et al. 1982) as
G

physical mechanisms including low-radio-power jets, star-

Sint bursts and shocks (e.g. Panessa et al. 2019). Owning
Tb = 1.22 × 109 2 2 (1 + z), (1)
RI

νobs θ size to the absence of bright radio cores and the missing
of sensitive short baselines, VLBI observations of the
where Sint is the integrated flux density in mJy, νobs is the radio counterparts of these galaxies often show non-
O

observation frequency in GHz, θ size is the FWHM of the cir- detections or significantly over-resolved radio structures.
cular Gaussian model in mas, and z is the redshift. The peak These cases involve extreme UV-selected starburst galax-
component C has a brightness temperature of ∼1 × 107 K. ies (Alexandroff et al. 2012), ULIRG IRAS 23365+3604
 Nuclear radio activity in PDS 456 5

Table 2. Summary of characteristic parameters of the radio components detected in PDS 456. Columns give (1) component name, (2)
total flux density, (3-4) relative offsets in Right Ascension and Declination with respect to the peak component C, (5) best-fit size θsize
of the Circular Gaussian model, (6) brightness temperature Tb and (7) radio luminosity LR .

Name Sint δra δdec θsize log Tb log LR

(mJy) (mas) (mas) (mas) (K) (erg s−1 )

C 2.31 ± 0.03 +0.0 ± 0.1 +0.0 ± 0.1 11.1 ± 0.4 6.99 ± 0.05 39.60 ± 0.04

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

NW 1.15 ± 0.03 −62.7 ± 0.2 +48.8 ± 0.2 14.7 ± 0.7 6.45 ± 0.06 39.30 ± 0.04
W 8.18 ± 0.14 −46.3 ± 0.9 −11.0 ± 0.8 118.6 ± 1.6 5.48 ± 0.05 40.15 ± 0.04

(Romero-Cañizales, Pérez-Torres & Alberdi 2012), hot detected based on the shortest baseline of Ef–Wb on the re-
dust-obscured galaxies (Frey et al. 2016) and sub- ceding side in the early observations at 5 GHz. Thus, the
millimetre-selected galaxies (Chen et al. 2020). component X1 is probably an instrumental noise peak.

PT
4.2.1 Location of the radio core 4.2.3 Diffuse component W: a relic jet or a composite of

RI
starbursts and shocks
The peak component C is the most promising component
hosting the radio core of PDS 456. As the only central fea- PDS 456 has an optically thin integrated spectrum up to

SC
ture in the radio emission distribution, it is quite close to >15 GHz (Yang et al. 2019). As the component W con-
the optical centroid (J2000, R.A. = 17h 28m 19.s 789380, Dec. = tributes about half of the total radio emission of PDS 456
−14◦ 15′ 55.′′ 85543, σp = 0.04 mas) reported by the Gaia as- and the components C has a relatively hard spectrum, the

U
trometry (DR2, Gaia Collaboration 2018) using a point- component W may have a similar optically thin spectrum
source model. The small offset (radius ∼4.5 mas) is very originating from non-thermal radio activity.

N
likely due to a double-peaked optical brightness distribu- It is hard to naturally identify the component W as a
tion (peak separation ∼0.22 kpc, Letawe, Letawe & Magain jet component extending further from the component NW,
as W has a diffuse radio morphology and a rather different
2010). The component C might have a relatively hard spec-
A
trum. Assuming no significant flux density variability be- extension direction from the existing jet, and there is no hint
of a bending point. With respect to the component C, the
M
tween the previous EVN observations at 5 GHz (Yang et al.
2019) and the new observations at 1.6 GHz, it has a spectral component W shows a very wide-angle (>180◦ ) extension
index of α = −0.5 ± 0.1. with a size of 118 ± 2 mas (366 ± 6 pc) and a centroid in
PA ∼ − 103◦ .
ED

The component W may be an extended relic jet compo-

nent close to the line of sight. Due to the peculiar jet geom-
4.2.2 Evidence of a low-radio-power jet
etry, the projected structure does not resemble an elongated
IT

There exists some faint emission smoothly connecting the jet. Compared to the inferred jet component NW, the identi-
components C and NW. The component NW could repre- fication necessitates a significant change in the jet direction.
sent the head of a jet along PA ∼ −52.◦ 1. With respect to the Currently, this information is unavailable.
ED

component C, the apparent jet opening angle for the compo- Because of its high luminosity and extended morphol-
nent NW is ∼10.◦ 7. This is about a factor of two smaller than ogy, it can not be simply explained as a single supernova or a
the median value of 21.◦ 5 found in significantly beamed AGN supernova remnant, which have typical peak monochromatic
jets (Pushkarev et al. 2017). Thus, the jet is unlikely to be luminosities only up to LR ∼1038.7 erg s−1 (Weiler et al.
N

very close to the line of sight. Assuming that the jet is close 2002) and might only explain the brightened spot in the
to the kinematic axis of the molecular disk (Bischetti et al. South. It is also hard to explain the component W as a
U

2019), the jet viewing angle is θ view = 25◦ ± 10◦ . This in- group of supernovae and supernova remnants, i.e. starbursts.
dicates a de-projected length of 570 ± 20 pc for the compo- The nearby extreme starburst galaxy Arp 220 has a star-
formation rate (SFR) of ∼220 M⊙ yr−1 and a radio luminos-
L

nent NW. Since there is no highly relativistic jet observed

among extreme accretion systems (Yang et al. 2020a), the ity of 4 × 1039 erg s−1 (Varenius et al. 2016, 2019). Compared
A

jet in PDS 456 might not be significantly Doppler boosted. to Arp 220, PDS 456 has at least three times lower SFR (cf.
Based on its low radio luminosity, it can be identified as a Sect. 4.1). Assuming that fifty per cent of radio emission is
IN

low radio power jet (e.g. Kunert-Bajraszewska et al. 2010; from the starburst activity, the SFR derived based on the ra-
An & Baan 2012). dio luminosity (Eq. 6, Bell 2003) would be 690 ± 70 M⊙ yr−1 .
The component NW has also been marginally detected This is inconsistent with the SED-based SFR and thus re-
G

as the component X2 in the previous EVN observations at quires an additional source of energy injection.
5 GHz (Yang et al. 2019). The two observations give a spec- Powerful high-speed AGN winds and outflows can
RI

tral index of α = −0.8 ± 0.1. Owing to the near face-on naturally produce the additional synchrotron radio emis-
disk geometry (Bischetti et al. 2019) and the southern dif- sion via shocks on scales ∼ > 100 pc in the nuclear high-
fuse emission, only the jet on the approaching side is clearly density star-forming region (e.g. Zakamska & Greene 2014;
O

identified. In the VLBI image at 1.6 GHz, we failed to con- Nims, Quataert & Faucher-Giguère 2015). Mildly relativis-
firm the existence of the component X1 (relative offsets, tic and wide-opening-angle winds have been identified in
δra ≈ 37.4 mas, δdec ≈ −99.4 mas), which was tentatively X-ray spectroscopic observations (e.g. Nardini et al. 2015;
6 J. Yang et al.
Matzeu et al. 2017; Reeves et al. 2020), with reports of mul- rate MÛ Edd , LEdd = ǫ MÛ Edd c2 where ǫ is the efficiency factor,
tiple velocity components, up to 0.46 c (Reeves et al. 2018). rbh = 6 rG = 6 GMbh /c2 (rG is the gravitational radius) is the
A possible ultraviolet outflow at 0.3 c has also been reported assumed distance from which the radiation is assumed to
by Hamann et al. (2018). Thus, mildly relativistic winds can peak and corresponds to the innermost stable circular orbit
extend farther out to pc scale (Reeves et al. 2020), as also for a Schwarzschild black hole (non-rotating). The evolution
inferred from Section 4.3 and thus might have a large im- of velocity of a free particle driven by radiation pressure from
pact on the whole nuclear region. Two CO(3-2) molecular the inner disk is given by (e.g. Abramowicz, Ellis & Lanza
outflow components were identified by Bischetti et al. (2019) 1990; Mohan & Mangalam 2015)

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

in the compact nuclear region. One is a blue-shifted, high- !
1/2
velocity (v ∼ −800 km s−1 outflow component. The other is a dβ (1 − β2 )
 
Γe ξs 2 8
< 500 km s−1 ), high-velocity dispersion (peak: = 1+β − β −1 , (4)
low-velocity ( ∼ dx 2x 2 βξ (1 − β2 )1/2 ξ 1/2 3
−1
360 km s ) outflow component, which is ∼50 mas west from
the quasar and spatially coincident with the centroid of the where x = r/rG , ξ = (1 − 1/x), ξs = ξ(x = 6) and we have
radio component W. The coincidence agrees with the shock assumed a radial particle trajectory which is shown to be

PT
model. the case for the large distance regime (Mohan & Mangalam
The latter scenario involving shocks due to wind in- 2015). The formulation corresponds to a wide-angle wind
teraction is now probed further to estimate the SFR. This outflow launched from the innermost region subject to a
radiative deceleration (an effective drag force), aided by the

RI
involves driving of galactic winds and consequent feedback
with the host galaxy through star formation. The relevant gravitational potential of the black hole. This can be used to
forces in action on cold gas clouds (temperature ≤ 104 K) at estimate the Eddington ratio corresponding to an observed

SC
pc–kpc scales include ram pressure from hot outflowing gas asymptotic wind velocity at the sub-pc – pc-scale.
(in which the cold clouds are embedded) and radiation pres- The equation (4) is solved assuming ǫ = 0.1 and Γe =
sure from the galactic disk which are in opposition to the 0.1 − 10.0 (sub – super Eddington luminosity) which corre-
sponds to mÛ = 0.12 − 12.0 using equation (3). For a Keple-

U
gravitational force due to the self-gravitating region of the
rial angular velocity Ωbh = GMbh /rbh 3 associated with the
disk (e.g. Sharma & Nath 2012); the corresponding SFR is
launch radius rbh , the rotational velocity vφ = rbh Ωbh =

N
approximated in terms of the wind velocity vw and the cir-
cular velocity of the galactic disk vc c/61/2 ≈ 0.41 c. This could represent the minimum veloc-
ity of disk material that is advected into the winds and
  5/2   25/8
A
vw vc
SFR = (0.89 M⊙ yr−1 ) . (2) possibly into the collimated jet; for the boundary condi-
vc 120 km s−1 tion associated with equation (4), we assume an initial
M
For the above estimated vw of 360–800 km s−1 and a fiducial launch velocity β(rbh ) = 0.41. A launch velocity exceed-
value of vc = 120 km s−1 , the SFR is 13.9–102.1 M⊙ yr−1 . ing 0.41 c could indicate a poloidal component away from
As this range is consistent with the estimated 69 M⊙ yr−1 the disk plane and hint at mechanisms driving the mate-
ED

from the IR observations, the latter is preferred. rial into trajectories away from the canonical Keplerian disk
If the component W is associated with the nuclear orbits, including a truncated disk scenario with a magnet-
molecular outflows, and formed by the AGN winds sweep- ically arrested disk accretion in the innermost region (e.g.
IT

ing up the high-density star-forming region, the energy con- Tchekhovskoy, Narayan & McKinney 2011) and energy ex-
version efficiency is LLbol
R
∼10−7 . This is two orders of mag- traction from the spin of the black hole (Blandford & Znajek
nitude lower than predicted by the model of spherical wind 1977; Blandford & Payne 1982).
ED

shocks (Nims, Quataert & Faucher-Giguère 2015). However, The above formulation can be generalized to estimate
this might be acceptable for PDS 456 as the nuclear winds the velocity (or Lorentz factor) in the launching region for
can preferentially propagate along the low-density polar di- sources which indicate accretion rates near or exceeding the
Eddington rate. The contours of the asymptotic velocity
N

rections. Our estimate is also consistent with the study of

Zakamska & Greene (2014) which finds a strong association β(Γe, x) are plotted in Fig. 3. The particle velocity tends to
between powerful outflows and the radio luminosity in radio a constant by ≈ 1000 rG ≈ 0.1 pc (in the AGN rest frame).
U

quiet quasars, with a statistical median conversion efficiency With an increase in Γe , the velocity saturates to a larger
of LLbol
R
∼ 10−6 . value for the assumed β(rbh ). Wind velocities of ≤ 0.3 c,
as suggested by the radio and X-ray observations can orig-
L

inate from the inner disk. These can be achieved well in

A

4.3 Wind velocity and the Eddington ratio advance and stay at the saturated value up to the pc-scale
If the powerful wide-angle X-ray winds (Nardini et al. 2015) and possibly beyond. The Eddington ratio corresponding to
IN

are radiatively driven (Matzeu et al. 2017), PDS 456 might β ≤ 0.3 c is Γe ≤ 2.6 ( MÛ ≤ 71.3 M⊙ yr−1 ), indicating that
have a super-Eddington accretion rate. The Eddington ratio PDS 456 likely accretes at rates exceeding the Eddington
Γe is the ratio of the accretion disk luminosity (Ldisk ) to the rate. The estimated Γe ≤ 2.6 suggests that the bolomet-
G

Eddington luminosity (LEdd ). For an optically thick geomet- ric luminosity (Reeves et al. 2000) may have been slightly
rically thick disk (Shakura & Sunyaev 1973) in the vicinity underestimated (upto 3 × 1047 erg s−1 ) or the SMBH mass
RI

of the SMBH, (Nardini et al. 2015) may have been overestimated (Mbh as
low as 3 × 108 M ⊙ ).
L 1 GMbh MÛ mÛ
Γe = disk ≈ = , (3) The formulation can act an independent manner of con-
O

LEdd LEdd 2rbh 12ǫ straining the Eddington ratio (and the associated accretion
where Mbh is the SMBH mass, MÛ is the accretion rate and rate) in this and similar super-Eddington sources using the
mÛ is the accretion rate scaled in terms of the Eddington observed wind velocity as an input.
 Nuclear radio activity in PDS 456 7
Contours of final velocity β (in units of c) Shandong Provincial Natural Science Foundation, China
5 (JQ201801). The European VLBI Network (EVN) is a
joint facility of independent European, African, Asian,
and North American radio astronomy institutes. Scientific
results from data presented in this publication are derived
4 from the following EVN project code(s): EY027. e-VLBI
research infrastructure in Europe is supported by the
European UnionâĂŹs Seventh Framework Programme

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

(FP7/2007-2013) under grant agreement number RI-261525
Eddington ratio

3
NEXPReS. e-MERLIN is a National Facility operated by
the University of Manchester at Jodrell Bank Observatory
on behalf of STFC. The research leading to these results has
received funding from the European Commission Horizon
2
2020 Research and Innovation Programme under grant

PT
agreement No. 730562 (RadioNet). This work has made use
of data from the European Space Agency (ESA) mission
Gaia (https://www.cosmos.esa.int/gaia), processed by
1
the Gaia Data Processing and Analysis Consortium (DPAC,

RI
https://www.cosmos.esa.int/web/gaia/dpac/consortium).
Funding for the DPAC has been provided by national insti-

SC
tutions, in particular the institutions participating in the
0
0 200 400 600 800 1000 Gaia Multilateral Agreement. This research has made use
r/ rG of the NASA/IPAC Extragalactic Database (NED), which
is operated by the Jet Propulsion Laboratory, California

U
Figure 3. Contours of velocity β from equation (4) as a function
Institute of Technology, under contract with the National
of the Eddington ratio and the distance from the SMBH r/rG (up
Aeronautics and Space Administration. This research

N
to 1000, about 0.1 pc). The β values tend to a constant before
about 0.1 pc for an initial launch velocity of 0.41 c. The radia- has made use of NASAâĂŹs Astrophysics Data System
Bibliographic Services.
tively driven and mildly relativistic winds with β = 0.1–0.3 c
A
indicate the Eddington ratio Γ = 0.2–2.6.
M

5 CONCLUSIONS DATA AVAILABILITY

The luminous radio-quiet quasar PDS 456 is a nearby ex- The correlation data of the EVN experiment EY027 are
ED

treme accretion and feedback system. With the EVN plus available to the public in the EVN data archive. The cal-
e-MERLIN observations at 1.66 GHz, we achieved a more ibrated visibility data and the final image files underlying
complete view of its complex radio morphology. Our VLBI this article will be shared on reasonable request to the cor-
responding author.
IT

images revealed a complex radio-emitting nucleus consist-

ing of a collimated jet and a very extended structure. The
latter has a size of about 120 mas and a radio luminosity
ED

about three times higher than the nearby extreme starburst REFERENCES
galaxy Arp 220. The powerful nuclear radio activity could
be explained as a relic jet with an unusual jet geometry, Abramowicz M. A., Ellis G. F. R., Lanza A., 1990, ApJ, 361, 470
Alexandroff R., et al., 2012, MNRAS, 423, 1325
N

or a more natural consequence of star-forming activity plus

An T., Baan W. A., 2012, ApJ, 760, 77
AGN-driven wind shocks. In addition, fainter nuclear radio
An T., Baan W. A., Wang J.-Y., Wang Y., Hong X.-Y., 2013,
activity and a sub-mJy flat-spectrum jet base may exist.
U

MNRAS, 434, 3487

The observed relativistic X-ray winds of ∼0.3 c in the inner Bell E. F., 2003, ApJ, 586, 794
sub-pc-scale region suggests that PDS 456 likely accretes Bischetti M., et al., 2019, A&A, 628, A118
L

exceeding the Eddington rate. Future multi-frequency deep Blandford R., Meier D., Readhead A., 2019, ARA&A, 57, 467
VLBI observations of PDS 456 can potentially provide more Blandford R. D., Payne D. G., 1982, MNRAS, 199, 883
A

details on the complex nuclear activity. Blandford R. D., Znajek R. L., 1977, MNRAS, 179, 433
Boissay-Malaquin R., Danehkar A., Marshall H. L., Nowak M A.,
IN

2019, ApJ, 873,29

Böttcher M., Harris D. E., Krawczynski H., 2012, Relativistic Jets
ACKNOWLEDGEMENTS from Active Galactic Nuclei. Wiley, Weinheim
G

This work was partly supported by the Square Kilometer Briggs D. S., 1995, BAAS, 27, 1444
Chabrier G., 2003, PASP, 115, 763
Array pre-research funding from the Ministry of Science and
Chen H., et al., 2020, A&A, 638, A113
RI

Technology of China (2018YFA0404603) and the Chinese

Condon J. J., Condon M. A., Gisler G., Puschell J. J., 1982, ApJ,
Academy of Sciences (CAS, No. 114231KYSB20170003). 252, 102
EN acknowledges financial support from the agreement Fan L., Han Y., Nikutta R., Drouart G., Knudsen K. K., 2016,
O

ASI-INAF n.2017-14-H.0. LF acknowledges the support ApJ, 823, 107

from National Natural Science Foundation of China Fan L., Jones S. F., Han Y., Knudsen K. K., 2017, PASP, 129,
(NSFC, Grant No. 11822303, 11773020 and 11421303) and 124101
8 J. Yang et al.
Fan L., Knudsen K. K., Han Y., Tan Q.-H., 2019, ApJ, 887, 74 Yang J., Paragi Z., An T., Baan W. A., Mohan P., Liu X., 2020,
Fiore F., et al., 2017, A&A, 601, A143 MNRAS, 494, 1744
Frey S., Paragi Z., Gabányi K. É., An T., 2016, MNRAS, 455, Yang J., Gurvits L. I., Paragi Z., Frey S., Conway J. E., Liu X.,
2058 Cui L., 2020, MNRAS, 495, L71
Gaia Collaboration, 2018, A&A, 616, A1 Yun M. S., Reddy N. A., Scoville N. Z., Frayer D. T., Robson
Giroletti M., Panessa F., Longinotti A. L., Krongold Y., E. I., Tilanus R. P. J., 2004, ApJ, 601, 723
Guainazzi M., Costantini E., Santos-Lleo M., 2017, A&A, 600, Zakamska N. L., Greene J. E., 2014, MNRAS, 442, 784
A87

Downloaded from https://academic.oup.com/mnras/advance-article/doi/10.1093/mnras/staa2445/5893321 by guest on 10 October 2020

Greisen E. W., 2003, in Heck A., ed., Astrophysics and Space Sci- This paper has been typeset from a TEX/LATEX file prepared by
ence Library, Vol. 285, Information Handling in Astronomy: the author.
Historical Vistas. Kluwer, Dordrecht, p. 109
Hamann F., Chartas G., Reeves J., Nardini E., 2018, MNRAS,
476, 943
Han Y., Han Z., 2012, ApJ, 749, 123
Han Y., Han Z., 2014, ApJS, 215, 2
Han Y., Han Z., 2019, ApJS, 240, 3

PT
Keimpema A., et al., 2015, Exp. Astron., 39, 259
Kettenis M., van Langevelde H. J., Reynolds C., Cotton B., 2006,
Astron. Data Anal. Software Syst. XV, 351, 497

RI
Kormendy J., Ho L. C., 2013, ARA&A, 51, 511
Kunert-Bajraszewska M., Gawroński M. P., Labiano A., Siemigi-
nowska A., 2010, MNRAS, 408, 2261

SC
Lamastra A., Menci N., Fiore F., Antonelli L. A., Colafrancesco
S., Guetta D., Stamerra A., 2017, A&A, 607, A18
Letawe Y., Letawe G., Magain, P., 2010, MNRAS, 403, 2088
Matzeu G. A., Reeves J. N., Braito V., Nardini E., McLaughlin

U
D. E., Lobban A. P., Tombesi F., Costa M. T., 2017, MNRAS,
472, L15

N
Mohan P., Mangalam A., 2015, ApJ, 805, 91
Nardini E., et al., 2015, Science, 347, 860
Nims J., Quataert E., Faucher-Giguère C.-A., 2015, MNRAS, 447,
A
3612
O’Brien P. T., Reeves J. N., Simpson C., Ward M. J., 2005, MN-
M
RAS, 360, L25
Panessa F., Baldi R. D., Laor A., Padovani P., Behar E., McHardy
I. 2019, Nat. Astron., 3, 387
ED

Petrov L., Kovalev Y. Y., Fomalont E. B., Gordon D., AJ, 2006,
131, 1872
Pushkarev A. B., Kovalev K. K., Lister M. L., Savolainen T.,
2017, MNRAS, 468, 4992
IT

Reeves J. N., O’Brien P. T., Vaughan S., Law-Green D., Ward

M., Simpson C., Pounds K. A., Edelson R., 2000, MNRAS,
312, L17
ED

Reeves J. N., Braito V., Nardini E., Lobban A. P., Matzeu G. A.,
Costa M. T., 2018, ApJ, 854, L8
Reeves J. N., Braito V., Chartas G., Hamann F., Laha S., Nardini
E., 2020, ApJ, 895, 37
N

Romero-Cañizales C., Pérez-Torres M. A., Alberdi A., 2012, 422,

510
U

Sharma M., Nath B. B., 2012, ApJ, 750, 55

Shakura N. I., Sunyaev R. A., 1973, A&A, 500, 33
Shepherd M. C., Pearson T. J., Taylor G. B., 1994, BAAS, 26,
L

987
Simpson C., Ward M., O’Brien P., Reeves J., 1999, MNRAS, 303,
A

L23
Tchekhovskoy A., Narayan R., McKinney J. C., 2011, MNRAS,
IN

418, L79
Tombesi F., Meléndez M., Veilleux S., Reeves J. N., González-
Alfonso E., Reynolds C. S., 2015, Nature, 519, 436
G

Tombesi F., 2016, Astron. Nachr., 337, 410

Torres C. A. O., Quast G. R., Coziol R., Jablonski F., de la Reza
R., Lépine J. R. D., Gregório-Hetem J., 1997, ApJ, 488, L19
RI

Varenius E., et al., 2016, A&A, 593, A86

Varenius E., et al., 2019, A&A, 623, A173
Weiler K. W., Panagia N., Montes M. J., Sramek R. A., 2002,
O

ARA&A, 40, 387

Yang J., An T., Zheng F., Baan W. A.; Paragi Z.; Mohan P.;
Zhang Z.; Liu X., 2019, MNRAS, 482, 1701