MNRAS 000, 1–16 (2023) Preprint 6 February 2023 Compiled using MNRAS LATEX style file v3.0
A low frequency sub-arcsecond view of powerful radio galaxies in
rich-cluster environments: 3C 34 and 3C320
V.H.Mahatma,1★ A.Basu,1 M. J. Hardcastle,2 L.K.Morabito,3,4 R. J. van Weeren5
1Thüringer Landessternwarte, Sternwarte 5, 07778 Tautenburg, Germany
2Centre for Astrophysics Research, Department of Physics, Astronomy and Mathematics, University of Hertfordshire, College Lane, Hatfield AL10 9AB, UK
3Centre for Extragalactic Astronomy, Department of Physics, Durham University, Durham DH1 3LE, UK
4Institute for Computational Cosmology, Department of Physics, University of Durham, South Road, Durham DH1 3LE, UK
5Leiden Observatory, Leiden University, PO Box 9513, 2300 RA Leiden, The Netherlands
Accepted 2023 February 01. Received 2023 February 01; in original form 2022 December 17
ABSTRACT
Models of radio galaxy physics have been primarily based on high frequency (>1GHz) observations of their jets, hotspots, and
lobes. Without highly resolved low frequency observations, which provide information on older plasma, our understanding of
the dynamics of radio galaxies and their interaction with their environment is limited. Here, we present the first sub-arcsecond
(0.3”) resolution images at 144MHz of two powerful radio galaxies situated in rich cluster environments, namely 3C 34 and
3C 320, using the International Low Frequency Array Telescope. We detect for the first time at low frequencies a plethora of
structures in these objects, including strikingly large filaments across the base of the lobes in both sources, which are spatially
associated with dense regions in the ambient medium. For 3C 34 we report a spectral flattening in the region of the central
filament, suggesting that the origin of the filaments is related to the presence of large-scale ordered magnetic fields. We also
report periodic total intensity and spectral index banding of diffuse emission in the eastern lobe, seen for the first time in radio
galaxy lobes. The hotspot complexes are resolved into multiple fragments of varying structure and spectral index; we discuss
the implications for particle acceleration and jet termination models. We find at most smooth gradients in the spectral behaviour
of the hotspot structure suggesting that particle acceleration, if present, may be occurring throughout the complex, in contrast to
simple models, but different jet termination models may apply to both sources.
Key words: radiation mechanisms: non-thermal – methods: observational – techniques: high angular resolution - galaxies:
active – galaxies: clusters: intracluster medium – galaxies: jets
1 INTRODUCTION
Deep radio observations of the jets, hotspots, and lobes of radio
galaxies (or radio-loud AGN) over the past few decades have given
us crucial insights into their energetics and dynamics. Particularly
with the advent of sensitive instruments capable of observing nearby
radio galaxies at sub-arcsecond resolution, primarily at frequencies
greater than 1 GHz (e.g. Karl G. Jansky Very Large Array and e-
MERLIN), it has been found that the synchrotron plasma in radio
galaxy lobes and hotspots exhibit complex structures that challenge
early models of radio galaxy evolution (e.g. Fernini et al. 1997;
Hardcastle et al. 1997; Leahy et al. 1997).
The ‘twin-exhaust’ or ‘beam’ models refined by Scheuer (1974)
and Blandford & Rees (1974) gave important qualitative understand-
ing on radio galaxy energetics and dynamics: energy is transported
from the central nucleus to hotspots by bipolar jets. The jet-headmust
decelerate (near the hotspots) to conserve momentum, since the am-
bient medium is denser than the jet itself. This causes material to flow
backward from the jet-head, forming the radio lobes which are, at the
very least, in pressure balance with the surrounding medium as they
★ E-mail: v.mahatma93@gmail.com
continue to expand. If the expansion of the lobes is supersonic, they
will drive a shock into the ambient medium. These particular studies
considered simplistic uniform environments, but further (semi-) an-
alytic models (e.g. Turner & Shabala 2015; Hardcastle 2018) refined
this picture to consider radio source environments, such as power-law
(Kaiser & Alexander 1997), 𝛽 model (Cavaliere & Fusco-Femiano
1978) and halo mass dependent atmospheres (Arnaud et al. 2010).
Numerical simulations which include more detailed physics such as
bulk relativistic flow, magnetic fields and spectral ageing have been
performed for many years (e.g. Turner & Shabala 2015; English et al.
2016; Yates-Jones et al. 2022), however the description of straight
jets, two hotspots either side of the central engine and a spheroidal
shell containing the radio source and its surrounding shocked gas
that is symmetrical about the jet axis have remained principally un-
changed.
Observationally, however, radio galaxies exhibit components that
are far from this simplistic picture. Aside from the fact that Fanaroff-
Riley type-I radio galaxies (FR-I; Fanaroff & Riley 1974) are mor-
phologically different (since they do not form bright hotspots at the
leading edge of the source), Fanaroff-Riley type II (FR-II; Fanaroff
& Riley 1974) radio galaxies can include, but are not limited to:
non-visible jets on at least one side of the source, compact knots
© 2023 The Authors
ar
X
iv
:2
30
2.
01
35
7v
1 
 [a
str
o-
ph
.G
A]
  2
 Fe
b 2
02
3
2 V. H. Mahatma
along the visible jet, jet bending, multiple compact features that may
all satisfy the defining criteria of a hotspot (e.g. Leahy et al. 1997),
radio lobes that are non-spheroidal, ‘wings’ and tails that tend to
develop at the base of the lobes and expand transversely away from
the jet axis (so-called X and Z-shaped radio galaxies; e.g. Cotton
et al. 2020; Müller et al. 2021; Gopal-Krishna & Dabhade 2022),
filamentation (e.g. Ramatsoku et al. 2020) and fan structures of very
low surface brightness adjacent to the lobes (e.g. 3C 351, 3C 263;
Bridle et al. 1994). Deep Very Large Array (VLA) observations at
sub-arcsecond resolution of the brightest radio galaxies (e.g. identi-
fied by the Third Cambridge Catalogue of radio sources, 3CR; Laing
et al. 1983) have detected these complex morphological features at
GHz frequencies for many years (e.g. Rudnick & Anderson 1990;
Bridle et al. 1994; Fernini et al. 1997; Hardcastle et al. 1997; Gilbert
et al. 2004; Mullin et al. 2006; Shulevski et al. 2019; Rudnick et al.
2022), but no empirical relationships between these features and the
overall physical properties of radio galaxies (e.g. their environments)
that can be adopted by dynamical models have been found.
While more refined models are currently needed to better de-
scribe the physics of jets, lobes and hotspots, we are still miss-
ing information on the detailed spectral structure of radio galaxies
down to low (∼ 100MHz) frequencies. The aforementioned studies
have been performed primarily at frequencies above 1GHz, where
sub-arcsecond imaging (corresponding to a resolution element of
∼ 0.5 kpc at 𝑧 = 0.1) was possiblewith past instruments. Since higher
energy synchrotron-emitting electrons radiate quicker than lower en-
ergy electrons, the oldest plasma in radio galaxy lobes is detected
only at the lowest frequencies (above the turnover frequency due to
self-absorption or above a low-frequency energy cut-off). Spectral
information over a large frequency range (between 100MHz and
8GHz, for example) at high resolution therefore allows us to distin-
guish between both old and freshly accelerated particle populations
in the lobes, for which we are missing understanding of at kpc-scale
resolution. Standard models predict that the oldest plasma collects at
the base of the lobes near the radio core, however recent studies dis-
covering the development of large-scale filaments (e.g. Ramatsoku
et al. 2020), wings (e.g. Cotton et al. 2020) and re-acceleration of the
oldest plasma (e.g. de Gasperin et al. 2017) has uncovered further
questions on their evolution, and on their ubiquity amongst the pop-
ulation. Low frequencies also provide the required short baselines
that probe structure on large angular scales, important for viewing
the large-scale lobe-environment interactions seen in recent low fre-
quency studies (e.g. Hardcastle et al. 2019).
The interpretation of filamentation and other structures also lacks
robustness without information on the ambient medium, to which
the dynamical evolution is strongly linked. Recent low frequency
(144-675MHz) observations of the radio galaxy NGC507 revealed
an S-shaped filament emanating from the radio lobes, suggested as
being introduced due to gas sloshing in its rich group environment
(Brienza et al. 2022). Filaments associated with radio galaxies are
becoming more frequently detected as instruments become more
sensitive, and their formation is suggested to be driven by complex
environmental interactions (e.g. Hardcastle et al. 2019; Rudnick et al.
2021; Knowles et al. 2021; Fanaroff et al. 2021; Riseley et al. 2022),
but further studies are needed to assert whether their formation is
ubiquitous in the radio galaxy population in rich environments. Such
interactions are also important for cluster science, as one of the unan-
swered fundamental questions concerns the origin of the large-scale
diffuse radio emission (radio haloes and relics) found in an increasing
number of deep observations of clusters (see review by van Weeren
et al. 2019). The most plausible explanation is the enrichment of the
medium with relativistic particles and magnetic fields from the lobes
of radio galaxies. It is therefore important to understand how radio
galaxy plasma couples to the cluster environment, the precise mech-
anisms of which are unknown. A low frequency study (combined
with higher frequencies) of radio galaxies in cluster environments,
with resolution that is sufficient to view the substructures of hotspots
and lobes, is required to address these questions.
The hotspot ‘structure’ (or hotspot ‘complex’; Leahy et al. 1997)
in radio galaxy lobes is also important to understand. Many radio
galaxies, when viewed at a physical resolution of the order 1 kpc,
display multiple compact, bright and flat-spectrum features at the
leading edge of a given lobe that meet the definition of a hotspot (e.g.
Laing 1982). Many questions arise from this fact: the classical mod-
els of FR-II jet evolution assume jets terminate at a single hotspot
where particle acceleration of the jet material occurs (Blandford &
Rees 1974). If there are multiple hotspots in radio galaxy lobes, are
they all associated with the beam termination, and how do they form,
evolve and contribute to the electron energy distribution of the radio
source as a whole? There is typically a variation in brightness, com-
pactness and spectral index between the multiple ‘hotspots’ (Bridle
et al. 1994), while there is evidence to suggest particle acceleration is
taking place inmore than one hotspot in a given lobe (Hardcastle et al.
2007). Although sites of particle acceleration can unambiguously be
identified using X-ray data that are well-described by synchrotron ra-
diation, low frequency radio data (of the order 100MHz) can identify
areas of flat-spectrum emission relative to steep-spectrum emission
which may give clues to acceleration and the flow of plasma between
hotspots.
In this paper, the first of a series of papers describing sub-
arcsecond resolution radio observations of the brightest radio galax-
ies at 144MHz using the International Low Frequency Array (here-
after LOFAR; van Haarlem et al. 2013), we report observations of
two rich cluster powerful radio galaxies: 3C 34 and 3C 320. Our aims
are to describe and understand:
• Their morphology seen at 144MHz at 0.3 arcsec resolutionwith
respect to higher frequency data;
• The nature, evolution, and fate of the oldest plasma in the sources
and their interaction with their cluster environments;
• The dynamics and origins of multiple hotspots and their asso-
ciation to jet termination models.
In section 2 we describe the target selection and data reduction pro-
cedures. In section 3 we describe our analysis of 3C 34 and 3C 320,
using LOFAR 144MHz data, higher frequency radio data and other
multi-wavelength data. We discuss our results in section 4 and sum-
marize our conclusions in section 5. Throughout the paper, we define
the spectral index 𝛼 in the sense 𝑆 ∝ 𝜈−𝛼. We use a ΛCDM cos-
mology in which 𝐻0 = 67.4 km s−1 Mpc−1, Ω𝑀 = 0.315 and
ΩΛ = 0.685 (Planck Collaboration et al. 2020).
2 DATA
2.1 Target selection
With sensitivity and image fidelity as a primary driver, LOFAR is
the most suitable array for a deep low frequency study — the high
band antennas (HBA), observing at a typical central frequency of
144MHz, are distributed such that it offers excellent 𝑢𝑣-coverage at
the shortest and longest baselines (with a largest angular scale of
∼70 arcsec after station combination, discussed in Section 2.4). The
LOFARTwoMetre Sky Survey (LoTSS; Shimwell et al. 2019, 2022),
an ongoing project to survey the entire northern sky at 6 arcsec res-
olution, processing only the Dutch array, achieves a median RMS
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 3
of ∼ 83 μJy beam−1. Use of the international stations, which are in-
cluded in observations of some fields and provide baselines up to
2000 km, can typically achieve sub-arcsecond (∼0.3) resolution at
a similar sensitivity (Morabito et al. 2022), and is ideal for our sci-
ence. We used the LoTSS1 Data Release 2 (LoTSS DR2; Shimwell
et al. 2022) to select 3CRR (Laing et al. 1983) sources – these are
the brightest (> 10 Jy at 178MHz) radio galaxies for which there
are a wealth of available ancillary data and have been studied for
many years (e.g. Bridle et al. 1994; Leahy et al. 1997; Hardcastle
et al. 1997) and offer complementary information to the LoTSS data.
From the 3CRR sources observed by LoTSS at the time of selec-
tion2, we selected sources having a suitable nearby delay calibrator.
We also needed to ensure the target is within a 1.25 degree radius of
the LoTSS pointing primary beam – the typical field of view (FOV)
of the HBA international stations has a 2.5 degree diameter. Unfortu-
nately, a number of LoTSS observations had been performed without
the international stations, which limited our selection. With a hand-
ful of 3C objects remaining, we selected 3C 34 and 3C 320 as two of
the brightest FR-II radio galaxies residing in cluster environments,
sufficiently large in angular size (> 50 beamwidths across their linear
extent at 0.3 arcsec), showing evidence of multiple hotspots, and hav-
ing available higher frequency radio data at the same resolution and
multi-wavelength data describing their surrounding environment.
2.2 3C 34
3C 34 is an extended FR-II radio galaxy with a 178-MHz luminosity
of 𝐿178 = 2.7× 1027WHz−1 (Laing et al. 1983), at a redshift of 𝑧 =
0.69 (Spinrad et al. 1985). It lies at the centre of a rich compact cluster,
with extended emission line gas from the host galaxy overlapping the
radio lobes (McCarthy et al. 1995). An [Oiii] emission line image
was reported where the emission is elongated along the jet axis, a
commonly found phenomenon for (𝑧 > 0.6) powerful radio galaxies
known as the alignment effect, suggested as describing jet-induced
star-formation (McCarthy 1987). Optical Hubble Space Telescope
(HST) images show that the host has diffuse morphology, possibly as
a result of the superposition of cluster-member galaxies (McCarthy
et al. 1995). Deeper HST imaging revealed the presence of a long
and narrow region of blue optical emission in the middle of the
western lobe, suggested to also describe jet-induced star-formation
in a cluster member (Best et al. 1997), and supported by observations
of stronger Faraday depolarization (between 1GHz and 5GHz) in this
region (Johnson et al. 1995). In the radio, both lobes are known to
have a double-hotspot structure, and a jet is claimed to have been
detected on both sides of the source (a jet is clearly detected on the
western lobe while jet knots appear on the eastern lobe; Johnson et al.
1995), and is therefore likely to be orientated close to the plane of
the sky. The back-flowing material in the lobes also have evidence
for a deflection at the base to form ‘wings’ (Johnson et al. 1995).
The consistent distribution of electric field vector angles suggest that
these wings are part of a large scale back-flow originating from the
hotspots, and in general they display high degrees of polarization
(Johnson et al. 1995). 3C 34 is therefore an exemplary source to
study jet-induced star-formation, large-scale radio wings, multiple
hotspot structures and jet knots.
1 https://lofar-surveys.org/dr2_release.html
2 The LoTSS survey was ∼ 30% complete at the time of selection
2.3 3C 320
3C320 is a classical FR-II radio galaxy with a radio luminosity of
𝐿178 = 3.3×1027WHz−1, at a redshift of 𝑧 = 0.342. It is identified as
a brightest cluster galaxy (BCG) in the centre of a rich X-ray-emitting
cluster (Massaro et al. 2013), where a large-scale shock surrounding
the lobes was interpreted from the X-ray emission. This shock was
analysed with deep 0.5-5 keV Chandra observations (Mahatma et al.
2020), while Vagshette et al. (2019) studied the large X-ray cavities
associated with the radio lobes. While the shocks are clear evidence
of AGN feedback, Mahatma et al. (2020) determined the jet power
(𝑄jet = 1.05×1038W) to be consistent with the cooling luminosity of
X-ray photons in the cluster (𝐿cool = 8.48×1037W), suggesting that
the radio source is sufficiently powerful to overcome radiative losses
in the intracluster medium (ICM). There is also a central bar of X-ray
emitting material between the lobes or cavities, indicating a dense
region of accreting gas. A complex hotspot structure in the eastern
lobe was also detected with the 1.5GHz e-MERLIN observations
presented by Mahatma et al. (2020), while the western lobe shows a
clear jetwith knots in their deep 5GHzVLAdata. TheX-ray emission
of the ICM shows an elongated spheroidal morphology in projection,
perhaps tracing a relatively recent merger. Due to its relatively small
angular size of 18 arcsec, past studies at radio frequencies are severely
limited for this source.
2.4 Data processing
2.4.1 LOFAR
3C 34 and 3C 320 were observed in LoTSS pointings P018+31 and
P233+35 under project codes LC6_015 and LC10_010 on the dates
of 2017 January 16 and 2018 September 14, respectively. Both point-
ings were observed for a total of 8 hours. We extracted the target data
from the Long Term Archive3 (LTA), as well as the 10-minute ob-
servations of the primary flux calibrators (3C 196 and 3C 295 for
3C 34 and 3C 320, respectively) for these pointings, observed just
before the target pointings. 231 sub-bands were extracted as individ-
ual measurement sets for all four data sets (two targets and two flux
calibrators) each with 16 channels and a 1 second integration time, in
the frequency range 120.3MHz to 167.7MHz (sub-bands at a higher
frequency are prone to considerable RFI), with a central frequency
of 144MHz. Both 3C 196 and 3C 295 are resolved for LOFAR with
the international array, so we used high resolution models fixed to
the Scaife & Heald (2012) flux density scale to ensure good solu-
tions. For the pointing containing 3C 34 a total of 74 LOFAR stations
were used (the Dutch array consists of 24 core stations that are each
divided into two sub-stations and 14 remote stations), including 12
international stations (non-Dutch). The pointing containing 3C 320
was observed with a similar number of stations, but with the addi-
tion of a station in Ireland, giving the Ireland-Poland baseline as the
longest baseline available to LoTSS observations (a 𝑢𝑣 distance of
∼ 2 × 106𝜆).
First, we ran PREFACTOR4 (de Gasperin et al. 2019) on the cali-
brator data which first flags (using AOFLAGGER; Offringa et al. 2012)
and averages the data, then sets the flux density scale and corrects the
calibrator data for the polarization alignment (between 𝑋𝑋 and 𝑌𝑌 ),
bandpass, clock and total electron content (TEC) for all stations. The
bandpasses for the international stations were typically a factor of 3
3 https://lta.lofar.eu/Lofar
4 https://github.com/lmorabit/prefactor
MNRAS 000, 1–16 (2023)
4 V. H. Mahatma
higher in amplitude than the Dutch stations, reflecting the higher sta-
tion gains as more tiles are used. These corrections are then applied
to the Dutch stations, and a phase-only calibration is performed on
those stations using the TGSS-ADR1 sky model (Intema et al. 2017).
The rotation measure for all the stations is also found, to help correct
for bulk changes in the TEC. The pre-determined corrections (clock,
bandpass, rotation measure) are then applied to the target.
From the data, which have at this point been only corrected for
direction-independent effects, we used the foundational calibration
strategy described by Morabito et al. (2022) to calibrate the inter-
national stations. We created two data sets (using DPPP; van Diepen
et al. 2018) which are phase-shifted to the most suitable5 nearby cal-
ibrators6 from the Long Baseline Calibrator Survey (LBCS; Jackson
et al. 2022) and to the target. In these data sets the core stations are
combined into a single super station to provide a sensitive reference
antenna (Morabito et al. 2022), and the data are averaged in frequency
and time, leading to a final frequency resolution of 48 kHz per chan-
nel and a time resolution of 8 sec. We derived phase and amplitude
calibration solutions for the delay calibrator following the method
and pipeline described by van Weeren et al. (2021). These solutions
were merged into a single file in the Hierarchical data format ver-
sion 5 (The HDF Group 2010) and applied to the target data. Since
these solutions are direction-dependent, the final calibration proce-
dure involved performing total electron content (TEC) and phase
corrections on the target, in an iterative self-calibration procedure
until convergence. In this step for the first self-calibration iteration
we used a skymodel for the target, using a Gaussianmodel formed by
applying pybdsf to the 5GHz 0.3 arcsecond VLA images of 3C 34
and 3C 320 (see full details below), to correct the astrometry and
help convergence.
The final full-bandwidth images were produced with WSCLEAN
(Offringa et al. 2014) on the corrected data sets, using multiscale
(Offringa & Smirnov 2017) and multi-frequency CLEANing, and are
shown on the top panels of Figures 1 and 3. The background RMS
noises in these images are 109 μJy beam−1 and 77 μJy beam−1, for
3C 34 and 3C 320, respectively. We achieve dynamic ranges of 1047
for 3C 34 (Figure 1) and 391 for 3C 320 (Figure 3).
2.4.2 VLA and e-MERLIN
We use archival 1.5GHz and 5GHz data to complement the LOFAR
data in this study. To approximately match the dense 𝑢𝑣-coverage
obtained by the LOFAR observations in the shortest spacings, we
combine multi-array data with the VLA. Due to the smaller angular
size of 3C320, we additionally use e-MERLIN data in combination
with the VLA at L band to provide . 0.3 arcsec resolution.
For 3C 34, we use A, AnB and C-array data at 5GHz, and A-array
data at 1.5GHz. Table 1 lists the observations we extracted from
the NRAO archive. We reduced the data sets in the standard manner
using AIPS. 3C 286 was observed as the primary flux calibrator for
the 1.5GHz data and for the A and AnB array data at 5GHz, and
3C 48 for the 5GHz C-array data, applying the Baars et al. (1977)
flux scale in the process. This was used to set the flux density of the
phase calibrators (0133+476 for the 1.5GHz data and the 5GHz A
and AnB array data). BL Lac was used as a phase calibrator for the
5GHz C-array data, allowing us to calibrate the complex gains and
transferring these to 3C 34. For the 5GHz C-array data, we followed
the original analysis of these data from Johnson et al. (1995) by
5 Most compact, bright and nearest to the target.
6 0.64 degrees and 0.49 degrees away from 3C 34 and 3C 320, respectively
allowing for an increase of 3.7% of the flux density of 3C 48 since
the Baars et al. (1977) measurement at the time of the observations.
No bandpass or polarization calibration was performed. The data set
was subsequently flagged for RFI using the CASA (McMullin et al.
2007) task rflag after calibration. The calibrated data for the target
3C34 were split and self-calibrated using CASA, with several rounds
of phase calibration until the solution interval equated integration
time (10 s). Self-calibration for the 5GHz AnB and C-array datasets
were initiated using a model using the self-calibrated A-array data, to
ensure astrometric accuracy between the datasets. The final images
were produced using the CLEAN algorithm on all four data sets. The
5GHz combined A, AnB and C array image is shown in Figure 2,
with LOFAR contours overlaid, with a total flux density of 0.40 Jy,
consistent with that of Johnson et al. (1995).
For 3C 320, being only 18 arcsec in angular size, we use VLA
A-array and B-array data at 6GHz, and VLA A-array data and e-
MERLIN data at 1.5GHz (calibration, combination and imaging
were discussed and presented by Mahatma et al. 2020, and we refer
the reader to their study). The 6GHzVLA image is shown in Figure 4
with LOFAR 144MHz contours overlaid.
2.4.3 Spectral data and image alignment
Herewe state themethods used to obtain resolved spectral indexmaps
for 3C 34 and 3C 320, using the higher frequency radio data discussed
in Section 2. Due to the lack of sub-arcsecond data at 1400MHz, we
produce a sub-arcsecond spectral index map only using 144MHz
and 5000MHz data for 3C 34, giving two-point spectral indices at
a resolution of 0.38 arcsecond to study the hotspots. For the diffuse
emission in 3C 34, we produce spectral index maps at 1.2 arcsecond
without the 5000MHz data due to the low signal-to-noise for the
steep-spectrum regions around the base of the lobes. For 3C 320,
given the high resolution e-MERLIN data at 1400MHz, we produce
0.2 arcsecond resolution maps across the frequency coverage (144,
1400MHz, 6000MHz), giving four-point spectral indices by splitting
the 6000MHz data. To mitigate the effects of different calibration
strategies introducing (sub)-pixel shifts in the astrometry, all the
maps were aligned by Gaussian-fitting of point-like sources, in our
case the hotspots, detected in each map (Harwood et al. 2019). After
aligning the maps, we used BRATS (Harwood et al. 2013) to produce
spectral index maps, using a 5𝜎 total intensity detection threshold at
each frequency.
As a caution, we note the uncertainties on the integrated flux den-
sities that affect our resolved spectral index maps as a systematic
offset. The data sets presented here are calibrated using different flux
density scales – the LOFAR data are on the Scaife & Heald (2012)
scale, the VLA data for 3C 34 are on the Baars et al. (1977) scale and
the VLA data for 3C 320 are on the Perley & Butler (2017) scale. To
estimate the magnitude of the flux density scale uncertainty intro-
duced by this, we compared the flux density scales for the primary
flux calibrators used in the data presented. For 3C 34, we compared
the Baars et al. (1977) scale to the Scaife & Heald (2012) scale at
144MHz for 3C 286, the flux calibrator for the VLA observations
at 1.5GHz and at 5GHz, finding a difference of 18.6% (higher in
the Baars et al. 1977 scale). For 3C 320, we compared the Perley &
Butler (2017) scale to the Scaife & Heald (2012) scale at 144MHz
for 3C 286, also the flux calibrator for the VLA higher frequency
observations, finding a difference of 13.12% (higher in the Perley &
Butler 2017 scale).
To cross-check the systematic flux density uncertainties on the
LOFAR data, we compared the measured source flux density with
that measured in the 6C survey (Hales et al. 1988, 1993) at 151MHz
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 5
1h10m16.00s17.00s18.00s19.00s20.00s21.00s
RA (J2000)
+31°47'00.0"
10.0"
20.0"
30.0"
40.0"
De
c 
(J2
00
0)
10"=72 kpc
300 1000 3000 10000
Total intensity ( Jy bm 1)
1h10m16.00s17.00s18.00s19.00s20.00s21.00s
RA (J2000)
+31°47'00.0"
10.0"
20.0"
30.0"
40.0"
De
c 
(J2
00
0)
10"=72 kpc
300 1000 3000 10000
Total intensity ( Jy bm 1)
central filament
vertical bridge
Object a
edge tail
jet knots
hotspot complex
diffuse tail
diffuse tail
host
Figure 1. Top: LOFAR 144MHz image of 3C 34 with a synthesized beam of 0.24×0.40 arcsec and a background RMS of 109 𝜇Jy beam−1. Bottom: LOFAR
144MHz image of 3C 34 with optical HST contours (blue) with the f785LP filter (𝜆 = 8620Å). Insets show zoom-ins of the eastern hotspot complex and of the
limb brightening either size of a vertical bridge crossing ‘Object a’, a region of jet-induced star-formation (Best et al. 1997). The zoom-ins are overlaid with
LOFAR contours (grey) from the same data, describing 50𝜎 multiplied by 2 to the power of integers from 1 to 10, chosen to highlight compact structures. Other
interesting features are annotated and discussed in the text. The LOFAR image beam size is shown in the top panel.
MNRAS 000, 1–16 (2023)
6 V. H. Mahatma
1h10m16.00s17.00s18.00s19.00s20.00s21.00s
RA (J2000)
+31°47'00.0"
10.0"
20.0"
30.0"
40.0"
De
c 
(J2
00
0)
10"=72 kpc
50 125 500 1250
Total intensity ( Jy bm 1)
Figure 2. 3C 34 VLA 5000 MHz image using the combined archival A-, AnB- and C-array data, with a synthesized beam of 0.39×0.34 arcsec and a background
RMS of 40𝜇 Jy beam−1. Contours are of the 144 MHz image, describing 3𝜎 plus 𝜎 multiplied by 1.5 to the power of integers from 1 to 20. The synthesized
beam shape of the VLA data are shown on the lower left.
Source Project Array Frequency (MHz) Date Exposure (min) Beam (arcsec) Flux density (Jy) RMS (𝜇Jy beam−1)
3C34 LC6_015 LOFAR 120-168 2017 Jan 16 480 0.24×0.40 18.76 109
AG247 VLA-A 1417-1662† 1987 Sept 11 556 1.21×0.97 1.48 139
AF213 VLA-A 4820-4870 1991 Jul 17 163 0.43×0.38 0.26 40
AP380 VLA-AnB 4820-4870 1991 Dec 22 80 1.40×1.11 0.40 42
AG247 VLA-C 4715-4965∗ 1988 Mar 29 47 6.87×5.14 0.40 125
3C320 LC10_010 LOFAR 120-168 2018 Sept 14 480 0.33×0.18 9.78 77
15A-420 VLA-A 1010-1970 2015 Aug 1 240 1.53×0.87 1.81 67
CY4223 e-MERLIN 1250-1700 2018 Mar 22 960 0.16×0.15 1.02 30
15A-420 VLA-A 3980-7900 2015 Aug 1 240 0.42×0.22 0.40 7
15A-420 VLA-B 3980-7900 2015 Feb 15 90 1.12×0.74 0.43 35
Table 1.Radio observations of 3C 34 and 3C 320 used in this study. †Simultaneous sub-bands at 1417MHz and 1662MHz were observed, each with a bandwidth
of 12.5MHz. ∗Simultaneous sub-bands at 4715MHz and 4965MHz were observed, each with a bandwidth of 50MHz. Note that beam sizes are dependent on
the weighting scheme used during imaging, with Briggs robust values between -1 and 0.5.
(calibrated using measurements of Cygnus A), scaled to 144MHz
assuming a spectral index of 0.5. With these comparisons, 3C 34 has
an uncertainty of 18% (higher in the LOFAR data), while 3C 320
has an integrated flux density uncertainty of 16% (higher in the
LOFAR data). These factors, while large, are likely dominated by
absolute flux density scale errors mentioned above, but also by im-
perfect beam models (which depend on a given LOFAR pointing)
during the processing, meaning that the relative gains between the
flux density calibrator and the target are not well determined. For the
LoTSS DR2 catalogue using only the Dutch array, Shimwell et al.
(2022) implement a procedure to align the flux density scale with
6C and NVSS, leading to an overall flux density uncertainty of less
than 10%, with a further potential error of 10% due to positional
variations in the LoTSS flux density scale, which also depend on
position from the pointing centre (Shimwell et al. 2022). Given that
these uncertainties are on the Dutch array only, and that the addition
of the international stations resulting in stronger ionospheric effects
act to exacerbate flux density errors, we suggest that these uncer-
tainties are not extreme, given the data. Flux density uncertainties
in the range 10-20% have been used for spectral analysis in recent
studies of 3C sources using the full international array of LOFAR
(Sweijen et al. 2022; Kappes et al. 2022). It is important to note that
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 7
24.00s24.50s25.00s25.50s26.00s26.50s15h31m27.00s
RA (J2000)
+35°33'24.0"
30.0"
36.0"
42.0"
48.0"
54.0"
De
c 
(J2
00
0)
10"=48 kpc
500 1250 5000 12500
Total intensity ( Jy bm 1)
15h31m24.00s24.50s25.00s25.50s26.00s26.50s27.00s
RA (J2000)
+35°33'24.0"
30.0"
36.0"
42.0"
48.0"
54.0"
De
c 
(J2
00
0)
10"=48 kpc
500 1250 5000 12500
Total intensity ( Jy bm 1)
filament
scythe2
scythe1
thermal shock boundary
ICM bar
Figure 3. Top: LOFAR 144MHz image of 3C 320 with the international array with a synthesized beam of 0.33×0.18 arcsec and a background RMS of 77 𝜇Jy
beam−1. Bottom: Chandra 0.5-5 keV X-ray contours of the hot gas ICM surrounding 3C 320 (teal), overlaid on the LOFAR 144 MHz image. Contours are at
3𝜎 plus 𝜎 to the power of integers from 1 to 20. The scythes (see inset), filament and the thermal shock boundary driven by 3C 320 are shown with arrows.
MNRAS 000, 1–16 (2023)
8 V. H. Mahatma
24.00s24.50s25.00s25.50s26.00s26.50s15h31m27.00s
RA (J2000)
+35°33'24.0"
30.0"
36.0"
42.0"
48.0"
54.0"
De
c 
(J2
00
0)
10"=48 kpc
30 100 300 1000
Total intensity ( Jy bm 1)
Figure 4. 3C 320 VLA 6 GHz image with the A and B array with a synthesized beam of 0.32×0.18 arcsec and a background RMS of 6 𝜇Jy beam−1, with
LOFAR 144 MHz contours overlaid. Contours are of the 144 MHz image, describing 3𝜎 plus 𝜎 multiplied by 1.5 to the power of integers from 1 to 20. Beam
shapes are shown on the lower left.
these integrated flux density errors are systematic, and they do not
apply to differences between points in highly resolved spectral index
maps, particularly with large gaps in frequency coverage in the data
presented. Nevertheless, and unless otherwise stated, we adopt these
errors (combined with measurement errors based on background
RMS) for the purposes of spectral index fitting, with the caveat that
all quoted values are uncertain by a systematic offset predominantly.
For the VLA L-band data for both 3C 34 and 3C 320, we adopt un-
certainties of 10% and 14%, respectively, based on measurements
using the Green Bank Telescope (GBT) at 1400 MHz, catalogued by
Kellermann et al. (1969), which themselves have flux density uncer-
tainties of 15% and are therefore consistent with our measurements.
For the VLA C-band data for 3C 34, we adopt an uncertainty of 5%
based on measurements using the GBT at 4830MHz (Griffith et al.
1990). For the VLA C-band data for 3C 320, we use the standard flux
density calibration uncertainty of ∼ 2% (Perley & Butler 2017). We
analyse the resulting spectral index maps for both sources in Section
3.3.
Object 𝑧 LAS (") 𝐿178 (W Hz−1)
3C34 0.691 49 2.7×1027
3C320 0.342 20 3.3×1027
Table 2. Redshift, largest angular size and 178 MHz radio luminosity mea-
sured in the 3CRR survey.
3 ANALYSIS
3.1 Radio morphology
3.1.1 3C34
In Figure 1 (top panel) we present the LOFAR 144 MHz image at
∼0.3 arcsec resolution, and with optical HST contours7 overlaid (bot-
tom panel). A plethora of compact and diffuse emission are observed
in the 144MHz data, but the most striking feature is the large-scale
filament across the base of the lobes spanning∼16 arcsec in projected
linear size (corresponding to ∼114 kpc at the distance of 3C 34), seen
in this source for the first time. Johnson et al. (1995) discovered ap-
parent wings at the base of the lobes, which are generally described as
large-scale backflows that expand transversely from the jet direction
at the base of the lobes in powerful sources, but no filament was seen
with the 1.2 arcsec angular resolution of their data at 1500MHz, nor
with the 0.3 arcsec angular resolution of their data at 5000MHz. At
144MHz at ∼ 0.3 arcsec resolution (see Figure 2 for a comparison
7 HST image courtesy of Phillip Best
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 9
with 5000MHz), this emission is resolved into a central filament and
diffuse tails, as annotated in the bottom panel of Figure 1. The diffuse
tails, extended in the North-South direction, could be an example of
old plasma being distributed into the ICM by the central filament.
The region of the central filament has been reported to have a high
degree (& 50% in some places) of polarization (Johnson et al. 1995)
at > 1GHz, indicating ordered magnetic fields which could belong to
the central filament. Unfortunately, techniques for robust polarization
calibration for LOFAR did not exist at the time of writing, limiting
our study to the total intensity at 144MHz.
Filamentary structures are seen throughout the source – the south-
ern edge of the eastern lobe shows tails which seem to follow a
trajectory starting from just south of the leading hotspot to the base
of the lobe (marked as ‘edge tail’ in the bottom panel). The west-
ern lobe also shows such structures throughout. The leading western
hotspot is connected with a faint tail (above ‘Object a’) in the north-
ern edge of the western lobe, which extends further downstream
towards the core. There is also a very faint small-scale tail (marked
as ‘vertical bridge’) that bridges this region and another high-surface
brightness region towards the south, while crossing ‘Object a’ (see
Section 3.2). It is clear that these filaments are produced in regions of
highest turbulence (the lobes have higher depolarization ratios with
distance from the hotspots; Johnson et al. 1995), but we discuss their
nature through environmental and spectral analysis in the proceeding
sections.
3.1.2 3C320
In Figure 3 (top panel) we present the 144 MHz LOFAR image of
3C320, at a resolution of ∼ 0.3 arcsec. The source is dominated by
diffuse emission at 144MHz, and the lobe material is seen to extend
transversely to the jet axis towards the core. It can be seen clearly that
a sharp filament exists along the base of the western lobe, while the
eastern lobe shows two relatively faint scythe-like structures. These
structures, labelled ‘scythe1’ and ‘scythe2’ in the bottom panel (see
inset), as well as the filament, are strikingly sharp and seem to flow
transverse to the jet axis (see Figure 4 to see the detected jet at 6000
MHz). We discuss the nature of these features further in Section 3.2.
3C 320 has been scarcely studied in the past, prohibiting a detailed
multi-wavelength analysis based on the literature. The only high res-
olution (sub-arcsec) study of this source was presented by Mahatma
et al. (2020) with VLA observations at 1500MHz and 6000MHz,
but the scythes and filament are only seen with the 144MHz data.
We also detect diffuse extension of material just east of the eastern
lobe, where the lobe boundary is expected to exist. This would be
expected under a model in which the jet termination occurs on the
front or back boundary of the lobe in projection, as suggested for
wide-angle tail radio galaxies (Hardcastle & Sakelliou 2004). We
note that this extension is seen at 1500MHz from the data presented
by Mahatma et al. (2020), but not at 6000MHz, consistent with real
steep-spectrum emission.
3.2 Environment
3.2.1 3C34
In the bottom panel of Figure 1we present the LOFAR image overlaid
with HST contours at 𝜆 = 8620Å from Best et al. (1997). The
host galaxy has spheroidal morphology and is surrounded by diffuse
optical emission (due to a superposition of cluster galaxies in the line
of sight, as suggested byMcCarthy et al. 1995). There is also another
large spheroidal cluster member to the north. However, there are no
signs such as tidal features in the optical images of an interaction
pre- or post-merger, suggesting that lobe-merger interactions are not
responsible for the morphology of the lobes, or for the stripping of
lobe material to form the diffuse tails in the north-south direction.
Noticeably the large central filament is curved, in projection,
around the host galaxy, presumably reflecting the galaxy halo which
has a higher density than the surrounding ICM. This would indicate
the origin of the filament is related to the high density regions in
the ambient medium – there is either a density jump in the ambient
medium encountered by the back-flowing plasma, causing a shock to
be developed, and the back-flowing aged particles are re-energized in
this shock front, and/or magnetic fields are particularly ordered and
enhanced around the dense region. This is discussed further, through
evidence of a spectral index flattening at the location of the filament,
in Section 3.3.
‘Object a’ is a thin strip seen in optical and far-infrared images,
interpreted as being a region of a jet-cloud interaction driving star-
formation in a galaxy (Best et al. 1997). However, as we discuss in
Section 4.1 below, this is not in line with the detected jet axis (see
Figure 2), which aligns with the southern hotspot in the western lobe
with the core and the jet knots in the eastern lobe. Moreover, there
is a vertical bridge (on scales of a few tens of kpc) that seems to
perpendicularly cross ‘Object a’ and bridges the radio limb bright-
ening either side (north-south) of ‘Object a’. This region, termed as
a depolarization silhouette by Johnson et al. (1995), exhibits strong
depolarization between 1 GHz and 5 GHz compared to its surround-
ings – defining the depolarization 𝐷𝑃216 as the ratio of the percent-
age polarization between 21 cm and 6 cm, Johnson et al. (1995) find
𝐷𝑃216 ∼ 0.1 directly at the location of ‘Object a’ and the vertical
bridge and 𝐷𝑃216 ∼ 0.5 in the immediate surroundings. This implies
that the depolarizing medium, likely to be a high thermal gas density
region, has suitable conditions for filament formation. A high ther-
mal gas densitywould also explain the limb-brightening either side of
the vertical bridge, if the back-flowing plasma naturally avoids high
density regions, consistent with our picture for the central filament.
3.2.2 3C320
In the bottom panel of Figure 3 we display the 144MHz LOFAR
image with 0.5–5 keV Chandra contours overlaid, taken from Ma-
hatma et al. (2020). A thermal ICM shock driven by the radio lobes
(Vagshette et al. 2019; Mahatma et al. 2020) is clearly seen to sur-
round the radio source. Interestingly, a high surface-brightness bar of
emission at the centre of the ICM (marked as ‘ICM bar’) aligns well
with ‘scythe2’ and the central filament seen across the base of the
lobes, asmarked in Figure 3. It is possible that back-flowing plasma is
disrupted by the presence of this high density region, or that ordered
magnetic fields play a role in enhancing the radio emission, as might
be the cases for the origin of the features seen in 3C 34. We discuss
the implications for the spatial coincidence between filamentation
and the dense regions at the centre of the ICM in Section 4.2.
3.3 Spectral structure & particle acceleration
3.3.1 3C34
3C 34 clearly has multiple compact features in both lobes. While
the exact definitions of hotspots and jet knots are debated, we use
the traditional terminology: hotspots are distinct compact structures
found at the edges of an FR-II source associated with jet termination,
whereas jet knots are compact features as part of a well-collimated
jet. Where multiple hotspots are detected in a single lobe, as is
MNRAS 000, 1–16 (2023)
10 V. H. Mahatma
ph
sh2
sh1 knots
b1
jet
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
Spectral index ( )
obj a
phflow
sh1
sh2
limb tail
jet
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
Spectral index ( )
Figure 5. Spectral index maps of the eastern (left) and western (right) hotspot structures of 3C34. Maps are of two-point spectral indices between 144MHz and
5000MHz, convolved to a circular 0.38 arcsec beam (black circle on the lower left). Total intensity contours of the 144MHz data are overlaid, starting at 20𝜎
(to emphasize the brighter and more compact regions), multiplied by 1.5 to the power of integers from 1 to 20.
particularly the case for 3C 34, we adopt the term ‘hotspot complex’
used by Leahy et al. (1997) to refer to a group of hotspots at the end
of a given lobe that is not associated with a jet knot. Following Laing
(1989), we define the ‘primary hotspot’ as the most compact hotspot,
and all other compact features in the hotspot complex are ‘secondary
hotspots’.
In Figure 5, we display two-point spectral index maps between
144MHz and 5000MHz of the eastern and western lobes, at an
angular resolution of 0.38 arcsec. In the eastern lobe (left panel)
there are many visible compact features evident through the LOFAR
total intensity contours (grey); the jet knots are resolved into four
components and least two bright hotspots at the eastern edge. We
note that, as suggested by Johnson et al. (1995) and Mullin et al.
(2006), the jet axis is most likely that which aligns the jet knots
in the eastern lobe with the jet that is well detected in the western
lobe (as marked in the figure). The spectral index map reveals that
there is a new region containing flat spectrum components – the
expected jet termination point contains a region with ∼ 3 point
sources (yellow in colour) with the flattest spectral index in the lobes
with 𝛼1505000 ≈ 0.65 ± 0.10. Since these are the most compact and
flat-spectrum components residing in the hotspot complex, and since
they lie at the expected jet termination point, we associate this region
with that containing the primary hotspot (labelled ‘ph’), although
clearly this region itself can be called a ‘primary hotspot complex’.
The secondary, brighter and larger hotspot just south of the
primary hotspot region (‘sh1’) has 𝛼1505000 ≈ 0.75 ± 0.06, and
the other secondary hotspot to the south-east of this (‘sh2’) has
𝛼1505000 ≈ 0.85 ± 0.06. A strikingly similar effect is seen for the west-
ern lobe on the right panel of Figure 5, with the most southern and
primary hotspot (‘ph’) having 𝛼1505000 ∼ 0.72 ± 0.06 and the sec-
ondary larger hotspot at the leading western edge (‘sh1’) having
𝛼1505000 ∼ 0.91± 0.06. As in the eastern lobe, ‘sh2’ in the eastern lobe
is slightly less compact and further downstream from ‘sh1’. There
is a region of flat spectrum emission between the ‘ph’ and ‘sh1’,
which we have called the ‘flow’ region, with 𝛼1505000 ∼ 0.69 ± 0.12,
which is flatter than the secondary hotspot and similar in spectral
indices with ‘ph’. The total intensity contours show that emission
extends north-west of ‘ph’ toward the secondary hotspots, implying
a flow of plasma. In both lobes, it is clear that if the newly found
primary hotspots are undergoing particle acceleration, they are not
restricted to one location in their primary hotspot complexes, and
that the spectral indices of secondary hotspots steepen with distance
from the primary hotspot. This has implications for the dynamics
of particle acceleration, jet-lobe interactions and the origin of the
secondary hotspots, which we discuss in Section 4.1.
Another curious aspect of 3C 34 is the periodic surface brightness
banding. As can be seen in the LOFAR image on the top panel of Fig-
ure 1, the eastern lobe shows periodic patches of surface brightness
increases and depressions from the hotspots to the core. In the top
panel of Figure 6, we display the two-point spectral index map at 1.2
arcsec resolution, using the 144MHz and 1400MHz data, overlaid
with the 144MHz total intensity contours at the same resolution.
The banding pattern (labelled ‘b1’, ‘b2’ and ‘b3’) is most promi-
nently seen in the 144MHz data, and it can be seen that the spectral
indices flatten in the surface brightness depressions (see grey LO-
FAR contours) between the banding. In the bottom panel of Figure 6
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 11
filament
b1 b2 b3
jet
0.6 0.8 1.0 1.2 1.4 1.6
Spectral index ( )
0 25 50 75 100 125
Distance (kpc)
0.04
0.06
0.08
0.10
F
lu
x
d
en
si
ty
[J
y
b
ea
m
−1
] b1 b2 b3
filament
0.9
1.0
1.1
1.2
1.3
1.4
S
p
ec
tr
al
in
d
ex
144 MHz flux density
1.4 GHz flux density (×10)
Spectral index
Figure 6. Top: Two-point spectral index map between 144 MHz and 1400
MHz for 3C 34, for the regions detected above 5𝜎, with a common beam
size of 1.2 arcsec. Contours are of the 144 MHz LOFAR image at 1.2 arcsec
resolution. Bottom: Radial line profile along the major axis of the eastern lobe
of 3C 34 in total intensity at 144MHz (blue), 1400MHz (grey) and spectral
index between 144MHz and 1400MHz (orange). Error bars describe the
measurement uncertainties. The locations of the banding regions and the
central filament are annotated.
we plot flux densities at 144MHz and at 1400MHz (multiplied by 10
for presentation purposes) and spectral index across a 1-D radial line
profile from b1 to the central filament (perpendicular to the centre
of the minor axis of the lobe). We include measurement uncertain-
ties only, using error propagation for the spectral indices, to reflect
more appropriate uncertainties on pixel to pixel variations of diffuse
lower surface brightness emission. It can clearly be seen that the low
frequency (blue line) banding represents regions of higher surface
brightness and steeper spectral index than the regions in between
the bands, which show a clear spectral flattening; this is inconsis-
tent with gradual spectral ageing from the hotspots toward the core.
The total intensity banding at b1 and b2 is less clear at 1400MHz,
but the depression between b2 and b3 and the steep decline at b3 is
clearly consistent with banding. The spectral flattening to the west
of b3 is particularly interesting, as the flattening is precisely at the
location of the central filament, as marked in Figure 6. Further down-
stream, towards the wing just north-west of the filament, the spectral
index steepens again to 𝛼1501400 >1.5. To our knowledge such patterns
have not been observed in total intensity for radio galaxies, although
banded rotation measure patterns in radio galaxies, associated with
magnetic field compression of an external magnetized medium has
been reported by Guidetti et al. (2011). Meanwhile, the spectral flat-
tening of the central filament may indicate enhanced magnetic fields
within the source, or a re-energisation of back-flowing material. We
discuss this further in Section 4.2.
3.3.2 3C320
3C320 is known to have a complex hotspot behaviour in the eastern
lobewhen viewed at sub-arcsecond resolution (Mahatma et al. 2020),
while the western lobe shows only one compact flat-spectrum region,
consistent with the findings of Bridle et al. (1994) that more complex
hotspot structures are favoured on the counter-jet side 8.
In Figure 7 we display the four-point spectral index map of 3C 320
between 144MHz and 6000MHz. For purposes of better displaying
the distinction between flat and steep spectral structure in the lobes,
we have saturated the colour scale for 𝛼 6 0.3 where the radio
core, which is undetected or is dimmer than the surrounding lobe
material in the 144MHzdata, shows an inverted spectrum. The jetted,
western lobe shows a single compact component in total intensity
with a spectral index (𝛼 = 0.50 ± 0.08), which we associate with
the primary hotspot (labelled ‘ph’ in Figure 7). This would require
the jet to bend slightly to the south in order to terminate at ‘ph’.
The most interesting aspect is that flat-spectrum emission is seen
ahead of the primary hotspot in the form of diffuse components,
labelled ‘d’. Recessed hotspots with lobe material further upstream
are common, representing the case where the jet interacts with the
front or back boundary of the lobe in projection. The upstream diffuse
material (‘d’) also has flat spectral indices (0.52 6 𝛼 6 0.57) but
with values consistent with the suggestion that the compact region ph
is the primary hotspot. There is a region at the northern edge of the
western lobe containing flat spectral indices with 𝛼 = 0.56 ± 0.23.
This feature is similar to the limb tail seen in the northern edge of
the western lobe of 3C 34, but it is not associated with any obvious
distinct structure in any of the total intensity maps of 3C 320.
The eastern lobe also shows a single compact flat spectrum com-
ponent with 𝛼 = 0.46 ± 0.08, which we associate with the primary
hotspot label ‘ph’. This hotspot seems to be associated with an arc
(labelled ‘arc’) of similarly flat-spectrum emission, akin to those
found for the northern hotspot of 3C 33 (Rudnick & Anderson 1990)
and the southern hotspot of 3C 445 (Prieto et al. 2002). We again see
a diffuse region of emission with 𝛼 = 0.55 ± 0.10, although this is
behind the primary hotspot, unlike the case for the western lobe. The
fact that the eastern hotspot is not recessed is consistent with it being
on the counter-jet side, as is commonly the case (Bridle et al. 1994).
The origin and nature of the diffuse flat-spectrum regions in both
lobes is unclear, but nevertheless give hints toward distributed par-
ticle acceleration, commonly associated with FR-I jets. We discuss
the implications of these results in Section 4.1.
Unfortunately, the scythes and filament are not detected at higher
frequency, prohibiting a detailed spectral analysis. However, the non-
detection of the scythes and filament at 1500MHz suggests that the
spectrum must be steep having 𝛼1441500 > 1.2 in the brighter regions
8 It should be noted that a jet is also detected on the western side, albeit
much fainter than the eastern side, and hence the jets are orientated close to
the plane of the sky, as per the case of 3C34.
MNRAS 000, 1–16 (2023)
12 V. H. Mahatma
d
d
knots
ph
arc
ph
0.4 0.6 0.8 1.0 1.2
Spectral index ( )
Figure 7. Four-point spectral index map between 150 MHz and 6000 MHz for 3C 320, for the regions detected above 5𝜎, with all three maps convolved to a
common beam size of 0.2 arcsec (black circle). Contours are of the VLA 6000MHz total intensity emission at the same resolution. The locations of high energy
emission discussed in the text are annotated. The jet axis is not marked due to a lack of a clear counter-jet detection, but is clearly seen on the jetted side.
closer to the host galaxy, and 𝛼1441500 > 1 in fainter regions farther
away.9 This clearly indicates that the synchrotron emitting plasma
in the filament are composed of relatively old cosmic ray electrons
(CREs), rather than newly accelerated CREs of the type expected
to exist at hotspots. Nevertheless, our new LOFAR image shows
evidence for clear interaction between the oldest plasma (> 25 Myr,
based on the spectral ageing analysis of higher frequency data by
Mahatma et al. 2020) and its surrounding environment, other than
the presence of large scale X-ray cavities coincident with the lobes
(Vagshette et al. 2019).
Using the lower limit on the spectral index of 1 in the filament,
and assuming a cylindrical geometry of the filaments having path
length (𝑙) similar to the projected-width on the plane of the sky
(𝑙 = 3 kpc for a width of 0.6 arcsec at the redshift of 3C 320), we
estimated the equipartition magnetic field strength . 15 μG by using
the revised equipartition formula (Beck & Krause 2005).10 This
provides a lower limit on the synchrotron+inverse-Compton cooling
lifetime & 50Myr at 144MHz in the filament, a factor of two from
the lower limit lifetime of 25Myr found from higher frequency data
(Mahatma et al. 2020).
9 Here we have assumed a 2𝜎 non-detection limit at 1.52GHz for an RMS
noise of 50 μJy beam−1.
10 Here we have used the ratio of the number densities of relativistic protons
to electrons 𝐾0 = 0 and a volume filling factor of 1.
4 DISCUSSION
4.1 Models for particle acceleration and jet termination
The addition of low frequency (144MHz) data has unveiled new
hotspot structures and other areas of flat-spectrum emission in 3C 34
and 3C 320, while the sub-arcsecond resolution maps have helped to
distinguish between compact and more diffuse structures. We note
that the distinction between sites of particle acceleration and ageing
is best given by evidence for optical and X-ray synchrotron emission
in hotspot complexes, while in the radio regime the highest available
frequencies ensures sensitivity to only the most recently accelerated
particles. Nevertheless, with our spectral index maps and with the
overall morphology seen at sub-arcsec resolution between 144MHz
and 6GHz we can make qualitative statements about jet termination
models for 3C 34 and 3C 320.
Flat-spectrum emission at the jet termination point is not restricted
to one location – distributed particle acceleration is suggested to oc-
cur in the hotspot complexes of radio galaxies, particularly in sec-
ondary hotspots (e.g. Hardcastle et al. 2007), and also on physically
larger scales in the jets of FR-I sources (e.g. Evans et al. 2005).
Using spectral index maps we discover new regions (that are not
clearly visible in total intensity but are detected at >10𝜎) that are di-
rectly associated with the jet termination in both lobes of 3C 34 (‘ph’
or primary hotspot complex; Figure 5), which are multi-component
and have sizes of ∼ 2 kpc. The secondary hotspots, having sizes
larger than 4 kpc, have a steepening with distance from the primary
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 13
hotspot complex. On the other hand, 3C 320 displays only one com-
pact component in both lobes, and flat-spectrum diffuse and arc-like
components (Figure 7) exist in the vicinity of the primary hotspot.
Competing theories exist for the formation of secondary hotspots:
models that propose that all hotspots are connected to the energy
supply include the ‘splatter-spot’ model (Williams & Gull 1985), in
which jet material flows out of the primary hotspot and terminates
again in the locations of secondary hotspots. Other models suggest
that the jet end-point moves along the lobe minor axis, such as the
‘dentist-drill’ (Scheuer 1982) and precessing jet (Cox et al. 1991)
models. The major difference in physical processes between these
models is that, in the case of the dentist-drill or precession mod-
els, secondary hotspots are remnants left behind by the motion of
the jet where particle acceleration should cease, causing a steep-
ening of their spectra at high frequencies. On the other hand, if
secondary hotspots continue to be fed by the energy supply (i.e. the
splatter-spot model), particle acceleration will continue in them, and
we expect flat-spectrum emission at high frequencies. Note that jet
precession is generally invoked as a prerequisite for the dentist-drill
model where the tip of the jet asymmetrically changes direction, but
jets are also known to abruptly change direction upon entering the
lobe (e.g. 3C 175 Kellermann & Owen 1988). For our purposes, we
reference both models to suggest the same jet termination model, as
we only distinguish between jet models describing hotspots that are
all connected with the energy supply and those that are not.
3C 34 displays secondary hotspots that have more aged plasma
with distance from the primary hotspot(s) – this provides strong
evidence for the dentist-drill or precessing jet models acting in this
system. Thiswas suggested byBest et al. (1997), through the absence,
with their data, of a clear termination point of the jet head in the
eastern lobe. Furthermore, this would also explain the jet-induced
optical strip ‘Object a’, which does not lie along the current jet
direction but may have been formed from the jet from a previous
precession angle. If the previous state of the jet was such that it
connected the core with ‘Object a’ and ‘sh1’ in a straight line (see
right panel of Figure 5), then this would require the jet to have
precessed by 6◦ (in projection) in the clockwise direction to its current
position (marked as ‘jet’ in Figure 5), consistent with precession
angles inferred from hotspot separations in 3C sources (Best et al.
1995). The gradual steepening of spectral index from the primary
hotspots, particularly seen in the ‘flow’ region in the western lobe,
gives a consistent picture. The splatter-spot model cannot be ruled
out, but would require extreme deflection of the jet flow, as it would
have to bend ∼ 90◦ to the south from ‘ph’ to ‘sh1’ in the eastern lobe.
Moreover, the splatter-spot model cannot be reconciled with our
spectral index maps (since it would require the secondary hotspots
to have similarly flat spectral indices as for the primary hotspots).
3C 320 does not display clear secondary hotspots, but rather dif-
fuse flat-spectrum regions in front (western lobe) and behind (eastern
lobe) the primary hotspot, as seen in Figure 7 (marked as ‘d’). Dif-
fuse and distributed particle acceleration, as is evident through X-ray
synchrotron emission, is usually seen in FR-I jets, but has also been
reported in some FR-II sources (e.g. 3C 390.3; Hardcastle et al. 2007
and 3C 33; Kraft et al. 2007) where the radio emission is often ei-
ther weak or not correlated with the X-ray morphology. With the
current data we cannot provide strong evidence for particle acceler-
ation in these regions due to a lack of high resolution X-ray data, but
nevertheless the spectral information favour the splatter-spot model
in which flat spectrum emission is evident in the primary hotspots
and the diffuse regions. While the dentist-drill or precession models
would be consistent with the radio morphology (due to the pres-
ence of diffuse flat-spectrum regions and the arc in the eastern lobe),
the spectral characteristics in the hotspot complexes would require
steepening. Rudnick & Anderson (1990) discussed these models to
describe the northern hotspot complex of 3C 33, also containing a
single hotspot and an arc connecting the hotspot on the outer edge
of the lobe. Without spectral information, they could not discrim-
inate between the splatter-spot and the dentist-drill models. In any
case, we cannot rule out ageing and structures that are not connected
with particle acceleration, and therefore discrimination against any
particular model, if the break frequency occurs beyond 6000MHz.
More robust spectral information in this regard would require higher
frequencies – observations of primary hotspots show that a spectral
cut-off tends to be present at millimetre wavelengths (∼ 30GHz).
We conclude that, due to the differences in morphology and spec-
tral structure between 3C 34 and 3C 320, powerful sources in cluster
environments can require different jet termination models to account
for their energetics. We confirm, along with many other studies, that
the working surface at the head of the jet is not static, but dynamic.
The hotspots in a given lobe can vary in size, shape, position and
spectral indices as a function of time, and has clearly been demon-
strated with numerical simulations for many years (e.g. Burns et al.
1991). The most favourable models, using the available data, are the
dentist-drill or precession models and the splatter-spot model, for
3C 34 and 3C 320 respectively. If this is true, further work must be
undertaken to understand why cluster-environment sources differ in
terms of their jet-lobe interactions.
4.2 Nature and origin of the oldest lobe plasma: filaments,
scythes and banding
At 144MHz and at 0.3 arcsec resolution, we have presented the first
high fidelity images of the radio sources 3C 34 and 3C 320 in order
to understand the structure of the oldest radio galaxy plasma in rich
environments. We have detected lobe banding in 3C 34, scythe-like
filamentation in 3C 320 and large-scale (> 40 kpc) filaments near the
base of the lobes in both sources, all of which have steep spectra
indicating their association with the oldest plasma.
For both sources, the nature of the central filaments is clearer
through the interaction with their environments. For 3C 34 the fila-
ment bends around the host galaxy, or the halo of the host galaxy, ev-
ident in the bottom panel of Figure 1. For 3C 320 the filament (along
with ‘scythe2’; Figure 3 bottom panel) lies at a surface brightness
jump toward the central dense bar in the ICM. For both sources, the
remarkable alignment of central filamentation with the dense mate-
rial in the surrounding medium (as also recently seen in NGC507;
Brienza et al. 2022) implies a likely physical association between the
non-thermal radio plasma and the surrounding thermal plasma. Simi-
lar situations have been observed recently for the cases ofNest200047
(Brienza et al. 2021) and Abell 2626 (Ignesti et al. 2020), where in
the former study the combination of buoyancy forces and cluster mo-
tions (e.g. gas ‘sloshing’) is suggested to drive material away from
the lobes and produce filamentation. The strikingly large and col-
limated nature of filaments at the bases of the lobes in 3C 34 and
3C 320 cannot solely be explained by this process since the lobe ma-
terial immediately surrounding the filaments would also be affected,
in a morphological sense, unlike in the aforementioned cases where
large-scale lobes do not surround the filaments.
Large-scale ordered and/or compressedmagnetic fields in the ICM
in the shape of the filaments can also contribute to their increased sur-
face brightnesses relative to their surroundings – magnetohydrody-
namic simulations have suggested that filaments trace enhancements
in the magnetic field distribution in radio galaxies (Huarte-Espinosa
et al. 2011; Hardcastle 2013). This would be qualitatively consistent
MNRAS 000, 1–16 (2023)
14 V. H. Mahatma
with the implication from Figure 6, showing a spectral flattening of
aged plasma by ∼ 20% (the peak spectral index of b3 is ∼ 1.4 while
that for the filament is ∼ 1.15) at the location of the filament seen
at 144MHz. This flattening, which deviates from simple models of
spectral ageing with a smooth steepening from hotspot to core, may
also indicate that the filament contains re-energized particles – while
the spectral index of the filament is significantly steeper than what is
expected for Fermi-I particle acceleration, this spectral behaviour is
analogous to the process of ‘gentle re-energisation’ of particles in-
troduced by de Gasperin et al. (2017). In this case, downstream from
the radio lobes, tails at very low frequencies (< 100MHz) can be ob-
served, with only a slight flattening of the spectral index between the
base of the lobes and the tails, and such mild flattening is associated
with inefficient particle acceleration that is just balanced by radiative
cooling. If such a process is occurring in 3C 34 and 3C 320, this has
major implications for radio plasma physics: with such a mild flat-
tening, an additional particle acceleration process other than diffuse
shock acceleration and Fermi-I (thought to occur in hotspots) may
be present that is particular to cluster and high density environments
surrounding radio galaxies. Moreover, this scenario alludes to the
existence of long-lived high energy cosmic rays that must exist in
the ICM as a seed population that are later re-accelerated by cluster
processes to form haloes and relics (see review by van Weeren et al.
2019). Clearly, the oldest plasma extends outwards from the fila-
ments in the north-south direction, thereby distributing plasma into
the peripheries of the ICM. Even without particle re-acceleration, it
is clear that magnetic field enhancement plays a role in developing
large-scale filaments. Deeper data at sub-arcsec resolutions over a
large frequency range that can detect these filaments are needed to
firmly establish these results. Nevertheless, our results show a clear
association and interaction between large-scale filamentation and the
densest regions of the ICM.
The inner parts of the low surface brightness tails were previously
identified as wings from a large-scale backflow (Johnson et al. 1995).
We see that these diffuse tails are almost perpendicular to the jet axis.
X-shaped radio galaxies, where the lobe structure is seen to extend
perpendicular to the jet axis (e.g. PKS 2014-55 Cotton et al. 2020)
have been observed formany years, but the origin of the perpendicular
wings is still a debate – rapid changes in the black hole spin direction
(Roberts et al. 2015), multiple pairs of jets (Bruni et al. 2020) and
jet deflection by the hot halo of the host galaxy (Cotton et al. 2020)
have all been suggested to explain such morphology. We prefer the
explanation given by Capetti et al. (2002), who found a striking
correlation between the axes of wings and those of host galaxies
of radio sources – the wings are preferentially directed along the
minor axis of their elliptical host galaxies, and there is a preference
towards high projected ellipticity. This naturally allows the backflow
to expand laterally due to the steeper pressure gradient in the ICM
perpendicular to the jet axis, and has been modelled with numerical
simulations (Leahy & Williams 1984; Capetti et al. 2002; Giri et al.
2022). The HST image in Figure 1 does not conclusively show an
elliptical morphology but rather a spheroidal morphology for the
host, which McCarthy et al. (1995) attribute to a superposition of
cluster members in the line of sight. Given further that the orientation
of the host could be such that its major axis is inclined to the line
of sight, we cannot conclusively reject the model of Capetti et al.
(2002).
The simulations of Giri et al. (2022) also predict that the turbu-
lent nature of the plasma in wings or diffuse tails will induce in situ
re-energisation of particles through local shocks, causing the wing
material to survive over longer time scales than what would be ex-
pected under a model of only radiative and adiabatic losses. More
sensitive observations at high resolution at frequencies greater than
1GHz that would detect the tails will allow for a highly resolved spec-
tral analysis of the diffuse tails to confirm this scenario, for the case
of 3C 34. It is also unclear whether the tails are physically associated
with the central filament (or if the scythes and filament are associ-
ated with the lateral lobe expansion in 3C 320), but if the filament
contains enhanced magnetic fields, this would provide a mechanism
for material to be transported along the filament and toward the ICM
outskirts – as would be needed to support theories for the formation
of diffuse cluster radio sources.
The ‘edge’ tail seen on the south side of the eastern lobe of 3C 34
(see Figure 1) is a newly seen structure and is seen to extend down-
stream towards the base of the lobes. A consistent explanation with
the aforementioned analysis that filaments and tails are shaped by the
gas distribution of the ambient medium can be made if the edge tail is
produced by lobe-environment interactions – Figure 1 (bottom panel,
see annotation) shows that the tail is highest in surface brightness at
a location where there is a ‘dent’ at the lobe southern edge. The HST
image shows no galaxy or source in this location, and any interaction
would most likely arise from clumps of hot gas in the ICM. To our
knowledge this is the first example of filamentation along the inner
edge of a radio galaxy lobe and deeper observations, particularly in
the X-ray, are required to give a view of the ICM gas distribution.
The nature of the lobe banding in the eastern lobe of 3C 34 (Figure
6) is particularly difficult to understand with the current data. This is
the first case in radio galaxy lobes, to our knowledge, where there are
periodic large (few tens of kpc) diffuse patches of high surface bright-
ness which are steep-spectrum, relative to their surroundings along
the jet axis. Guidetti et al. (2011) reported banded Rotation Mea-
sure (𝑅𝑀) structure in the lobes of a small sample of radio galaxies,
suggesting that the corresponding Faraday rotation occurs due to the
lobes expanding into surrounding magnetized plasma, compressing
and enhancing the magnetic fields in the immediate surroundings.
Radio lobes are known to be in contact or to interact with their envi-
ronment – the presence of X-ray cavities seen in a number of objects
situated in clusters (e.g. McNamara & Nulsen 2007) suggests radio
galaxy lobes in general can displace ambient thermal gas. In the case
of 3C 34 the banded patches have a steeper spectrum than their sur-
roundings (see Figure 6), requiring lower magnetic field strengths if
the intrinsic electron spectrum is curved in this region, inconsistent
with the field compression scenario. However, the regions between
the bands do have flatter spectral indices (see ‘b1’ in Figure 5), which
may indicate regions of magnetic field compression, but such narrow
linear features expanding into the magnetized ICM quicker than the
surrounding lobe material is implausible.
The radio galaxy Hercules A has been observed to show spec-
tral differences between prominent ring structures and the surround-
ing lobe material in the western lobe, suggested as being associ-
ated with cyclic activity (Timmerman et al. 2022), and such sharp
spectral boundaries have been used to suggest cyclic jet activity in
other sources (e.g. 3C 388; Roettiger et al. 1994). 3C 34 on the other
hand does not show any prominent and sharp features that mark the
boundary of the bands of diffuse steep-spectrum material. A more
convincing explanation for the banding is that the backflow is pe-
riodic (rather than the jet activity) driven by jet precession, which
we have associated to the jets of 3C 34 due to the spectral properties
of the hotspots. For a precessing jet, particularly if the precession
properties change with time (simulations suggest wider precession
cone opening angles lead to slower growth, e.g. Horton et al. 2020),
the forward expansion rate is likely to be periodic. Therefore, the
backflow is likely to be periodic, leading to a pattern of long and
short-lived electron populations across the lobes toward the core. By
MNRAS 000, 1–16 (2023)
A low frequency sub-arcsec view of 3C 34 and 3C 320 15
calibrating radio source evolution models with radio observations,
Turner et al. (2018) determine the age of 3C 34 to be 25.1Myr. If
there are indeed three bands of distinct electron populations driven
solely by jet precession, then this constrains a potential precession
period of 8Myr, consistent with precession periods of 3C sources (of
the order of 1–10 Myr) based on the analysis by Krause et al. (2019).
Nevertheless, we can conclude that one of either standard dynamics
or standard shock acceleration mechanisms does not apply in the for-
mation of the banding. Further studies of other cluster-centre sources
at low frequencies will shed light on the ubiquity of such structures
and will help to determine empirical trends that challenge standard
models of radio galaxy evolution.
5 CONCLUSIONS
Our high resolution imaging of 3C 34 and 3C 320 at 144MHz has
revealed the complexities associated with both the highest and lowest
energy radio emission observed in the lobe and hotspot plasma. In
summary, our conclusions are:
• Large-scale filamentation spanning the width of the lobes of
powerful radio galaxies may be a common phenomenon in cluster
environments. The precise shape and location of the filaments is
associated with the density profile of the ambient environment.
• For 3C 34 we find a spectral flattening at the location of the
filament, suggesting that enhanced magnetic fields are responsible
for its formation. The resulting particle populations, that may be re-
energized, continue to be driven perpendicular to the jet axis to form
low surface brightness tails that can mix with the ICM, and provides
support to radio galaxy lobes providing the seed particles for the
existence of diffuse cluster sources.
• We report the first case of large-scale periodic banding in total
intensity and spectral index in the lobes of radio galaxies, evident in
the eastern lobe of 3C 34. The origin of these structures is unknown,
but may be a result of jet precession.
• The dynamic hotspot structure in the lobes reveal that the jet
termination point is fragmented into multiple closely situated regions
of very compact regions of similar spectral indices, but also diffuse
regions of flat spectral index in 3C 320.We argue that the dentist-drill
or precessing models and the splatter-spot model best describe the
jet behaviours in 3C 34 and 3C 320, respectively.
ACKNOWLEDGEMENTS
We thank the anonymous referee for helpful insights that has im-
proved the paper. We also thank Andrea Botteon and Chris Riseley
for helpful comments that improved the paper.
MJH acknowledges support from the UK STFC [ST/V000624/1].
This work was supported by the Medical Research Council
[MR/T042842/1]. RJvW acknowledges support from the ERC Start-
ing Grant ClusterWeb 804208.
This research made use of the University of Hertfordshire high-
performance computing facility and the LOFAR-UK computing fa-
cility located at the University of Hertfordshire (https://uhhpc.
herts.ac.uk) and supported by STFC [ST/P000096/1].
LOFAR is the Low Frequency Array, designed and constructed
by ASTRON. It has observing, data processing, and data storage
facilities in several countries, which are owned by various par-
ties (each with their own funding sources), and which are collec-
tively operated by the ILT foundation under a joint scientific policy.
The ILT resources have benefited from the following recent major
funding sources: CNRS-INSU, Observatoire de Paris and Univer-
sité d’Orléans, France; BMBF, MIWF-NRW, MPG, Germany; Sci-
ence Foundation Ireland (SFI), Department of Business, Enterprise
and Innovation (DBEI), Ireland; NWO, The Netherlands; The Sci-
ence and Technology Facilities Council, UK; Ministry of Science
and Higher Education, Poland; The Istituto Nazionale di Astrofisica
(INAF), Italy.
The National Radio Astronomy Observatory is a facility of the
National Science Foundation operated under cooperative agreement
by Associated Universities, Inc.
This researchmade use ofAPLpy, an open-source plotting package
for Python (Robitaille & Bressert 2012).
DATA AVAILABILITY
The data underlying this article will be shared on reasonable request
to the corresponding author.
REFERENCES
Arnaud M., Pratt G. W., Piffaretti R., Böhringer H., Croston J. H., Pointe-
couteau E., 2010, A&A, 517, A92
Baars J. W. M., Genzel R., Pauliny-Toth I. I. K., Witzel A., 1977, A&A, 61,
99
Beck R., Krause M., 2005, Astronomische Nachrichten, 326, 414
Best P. N., Bailer D. M., Longair M. S., Riley J. M., 1995, MNRAS, 275,
1171
Best P. N., Longair M. S., Rottgering H. J. A., 1997, MNRAS, 286, 785
Blandford R. D., Rees M. J., 1974, MNRAS, 169, 395
Bridle A. H., Hough D. H., Lonsdale C. J., Burns J. O., Laing R. A., 1994,
AJ, 108, 766
Brienza M., et al., 2021, Nature Astronomy,
Brienza M., et al., 2022, arXiv e-prints, p. arXiv:2201.04591
Bruni G., et al., 2020, MNRAS, 494, 902
Burns J. O., Norman M. L., Clarke D. A., 1991, Science, 253, 522
Capetti A., Zamfir S., Rossi P., Bodo G., Zanni C., Massaglia S., 2002, A&A,
394, 39
Cavaliere A., Fusco-Femiano R., 1978, A&A, 70, 677
Cotton W. D., et al., 2020, MNRAS, 495, 1271
Cox C. I., Gull S. F., Scheuer P. A. G., 1991, MNRAS, 252, 558
English W., Hardcastle M. J., Krause M. G. H., 2016, MNRAS, 461, 2025
Evans D. A., Hardcastle M. J., Croston J. H., Worrall D. M., Birkinshaw M.,
2005, MNRAS, 359, 363
Fanaroff B. L., Riley J. M., 1974, MNRAS, 167, 31P
Fanaroff B., et al., 2021, MNRAS, 505, 6003
Fernini I., Burns J. O., Perley R. A., 1997, AJ, 114, 2292
Gilbert G. M., Riley J. M., Hardcastle M. J., Croston J. H., Pooley G. G.,
Alexander P., 2004, MNRAS, 351, 845
Giri G., Vaidya B., Rossi P., Bodo G., Mukherjee D., Mignone A., 2022,
arXiv e-prints, p. arXiv:2203.01347
Gopal-Krishna Dabhade P., 2022, arXiv e-prints, p. arXiv:2207.03814
Griffith M., Langston G., Heflin M., Conner S., Lehar J., Burke B., 1990,
ApJS, 74, 129
Guidetti D., Laing R. A., Bridle A. H., Parma P., Gregorini L., 2011, Monthly
Notices of the Royal Astronomical Society, 413, 2525
Hales S. E. G., Baldwin J. E., Warner P. J., 1988, MNRAS, 234, 919
Hales S. E. G., Baldwin J. E., Warner P. J., 1993, MNRAS, 263, 25
Hardcastle M. J., 2013, MNRAS, 433, 3364
Hardcastle M. J., 2018, MNRAS, 475, 2768
Hardcastle M. J., Sakelliou I., 2004, MNRAS, 349, 560
Hardcastle M. J., Alexander P., Pooley G. G., Riley J. M., 1997, MNRAS,
288, 859
Hardcastle M. J., Croston J. H., Kraft R. P., 2007, ApJ, 669, 893
Hardcastle M. J., et al., 2019, MNRAS, 488, 3416
MNRAS 000, 1–16 (2023)
16 V. H. Mahatma
Harwood J. J., Hardcastle M. J., Croston J. H., Goodger J. L., 2013, MNRAS,
435, 3353
Harwood J. J., Vernstrom T., Stroe A., 2019, MNRAS, p. 2670
Horton M. A., Krause M. G. H., Hardcastle M. J., 2020, MNRAS, 499, 5765
Huarte-Espinosa M., Krause M., Alexander P., 2011, MNRAS, 418, 1621
Ignesti A., et al., 2020, A&A, 643, A172
Intema H. T., Jagannathan P., Mooley K. P., Frail D. A., 2017, A&A, 598,
A78
Jackson N., et al., 2022, A&A, 658, A2
Johnson R. A., Leahy J. P., Garrington S. T., 1995, MNRAS, 273, 877
Kaiser C. R., Alexander P., 1997, MNRAS, 286, 215
Kappes A., et al., 2022, A&A, 663, A44
Kellermann K. I., Owen F. N., 1988, in Kellermann K. I., Verschuur G. L.,
eds, , Galactic and Extragalactic Radio Astronomy. pp 563–602
Kellermann K. I., Pauliny-Toth I. I. K., Williams P. J. S., 1969, ApJ, 157, 1
Knowles K., et al., 2021, arXiv e-prints, p. arXiv:2111.05673
Kraft R. P., Birkinshaw M., Hardcastle M. J., Evans D. A., Croston J. H.,
Worrall D. M., Murray S. S., 2007, ApJ, 659, 1008
Krause M. G. H., et al., 2019, MNRAS, 482, 240
Laing R., 1982, in Heeschen D. S., Wade C. M., eds, Vol. 97, Extragalactic
Radio Sources. pp 161–162
Laing R., 1989, in Meisenheimer K., Roeser H.-J., eds, , Vol. 327, Hot Spots
in Extragalactic Radio Sources. p. 27, doi:10.1007/BFb0036010
Laing R. A., Riley J. M., Longair M. S., 1983, MNRAS, 204, 151
Leahy J. P., Williams A. G., 1984, MNRAS, 210, 929
Leahy J. P., Black A. R. S., Dennett-Thorpe J., Hardcastle M. J., Komissarov
S., Perley R. A., Riley J. M., Scheuer P. A. G., 1997, MNRAS, 291, 20
Mahatma V. H., Hardcastle M. J., Croston J. H., Harwood J., Ineson J.,
Moldon J., 2020, MNRAS, 491, 5015
Massaro F., Harris D. E., Tremblay G. R., Liuzzo E., Bonafede A., Paggi A.,
2013, ApJS, 206, 7
McCarthy P. J., 1987, in Bulletin of the American Astronomical Society.
p. 1068
McCarthy P. J., Spinrad H., van Breugel W., 1995, ApJS, 99, 27
McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007, in Shaw
R. A., Hill F., Bell D. J., eds, Astronomical Society of the Pacific Confer-
ence Series Vol. 376, Astronomical Data Analysis Software and Systems
XVI. p. 127
McNamara B. R., Nulsen P. E. J., 2007, ARA&A, 45, 117
Morabito L. K., et al., 2022, A&A, 658, A1
Müller A., et al., 2021, MNRAS, 508, 5326
Mullin L. M., Hardcastle M. J., Riley J. M., 2006, MNRAS, 372, 113
Offringa A. R., Smirnov O., 2017, MNRAS, 471, 301
Offringa A. R., van de Gronde J. J., Roerdink J. B. T. M., 2012, A&A, 539,
A95
Offringa A. R., McKinley B., Hurley-Walker et al., 2014, MNRAS, 444, 606
Perley R. A., Butler B. J., 2017, ApJS, 230, 7
Planck Collaboration et al., 2020, A&A, 641, A6
Prieto M. A., Brunetti G., Mack K.-H., 2002, Science, 298, 193
Ramatsoku M., et al., 2020, A&A, 636, L1
Riseley C. J., et al., 2022, MNRAS, 515, 1871
Roberts D. H., Cohen J. P., Lu J., Saripalli L., Subrahmanyan R., 2015, ApJS,
220, 7
Robitaille T., Bressert E., 2012, APLpy: Astronomical Plotting Library
in Python, Astrophysics Source Code Library, record ascl:1208.017
(ascl:1208.017)
Roettiger K., Burns J. O., Clarke D. A., Christiansen W. A., 1994, ApJ, 421,
L23
Rudnick L., Anderson M., 1990, ApJ, 355, 427
Rudnick L., Cotton W., Knowles K., Kolokythas K., 2021, Galaxies, 9, 81
Rudnick L., et al., 2022, arXiv e-prints, p. arXiv:2206.14319
Scaife A. M. M., Heald G. H., 2012, MNRAS, 423, L30
Scheuer P. A. G., 1974, MNRAS, 166, 513
Scheuer P. A. G., 1982, in Heeschen D. S., Wade C. M., eds, Vol. 97, Extra-
galactic Radio Sources. pp 163–165
Shimwell T. W., et al., 2019, A&A, 622, A1
Shimwell T. W., et al., 2022, A&A, 659, A1
Shulevski A., et al., 2019, A&A, 628, A69
Spinrad H., Djorgovski S., Marr J., Aguilar L., 1985, PASP, 97, 932
Sweijen F., et al., 2022, A&A, 658, A3
The HDF Group 2000-2010, Hierarchical Data Format, version 5
Timmerman R., et al., 2022, A&A, 658, A5
Turner R. J., Shabala S. S., 2015, ApJ, 806, 59
Turner R. J., Shabala S. S., Krause M. G. H., 2018, MNRAS, 474, 3361
Vagshette N. D., Naik S., Patil M. K., 2019, MNRAS, 485, 1981
Williams A. G., Gull S. F., 1985, Nature, 313, 34
Yates-Jones P. M., Turner R. J., Shabala S. S., Krause M. G. H., 2022,
MNRAS, 511, 5225
de Gasperin F., et al., 2017, Science Advances, 3, e1701634
de Gasperin F., et al., 2019, A&A, 622, A5
van Diepen G., Dijkema T. J., Offringa A., 2018, DPPP: Default Pre-
Processing Pipeline (ascl:1804.003)
van Haarlem M. P., et al., 2013, A&A, 556, A2
vanWeeren R. J., de Gasperin F., Akamatsu H., Brüggen M., Feretti L., Kang
H., Stroe A., Zandanel F., 2019, Space Sci. Rev., 215, 16
van Weeren R. J., et al., 2021, A&A, 651, A115
This paper has been typeset from a TEX/LATEX file prepared by the author.
MNRAS 000, 1–16 (2023)