1. Introduction
It is well known for several decades that the double radio sources hosted by early-type galaxies located inside galaxy clusters follow a significantly different evolutionary track as compared to those located outside the cluster environment (McNamara & Nulsen Reference McNamara and Nulsen2007; Fabian Reference Fabian2012; Hardcastle & Krause Reference Hardcastle and Krause2013; Pasini et al. Reference Pasini, Finoguenov, Brüggen, Gaspari, de Gasperin and Gozaliasl2021). The difference arises mainly due to the higher ambient density, temperature (and the resulting static pressure as well as sound speed) encountered by the relativistic plasma jets and the radio lobes inflated by them within the hot intra-cluster medium (ICM). Such interactions quicken the deceleration of the jets and weaken the associated Mach disks, leading to morphological transformation/deformation of the radio lobes. Secondly, a more efficient confinement of the lobes by the ICM static pressure delays their fading due to adiabatic expansion, thus allowing the synchrotron losses to accumulate and manifest as an ultra-steep spectrum (USS) of the radio lobes (Baldwin & Scott Reference Baldwin and Scott1973). Consequently, the USS radio lobes can be detected at low radio frequencies long after the cessation of energy injection by the jets, whereafter the lobes can rise within the ICM due to its buoyancy pressure (Gull & Northover Reference Gull and Northover1973). Such faded lobes are sometimes identified as X-ray ‘cavities’ (Böhringer et al. Reference Böhringer, Voges, Fabian, Edge and Neumann1993; McNamara et al. Reference McNamara, Nulsen, Wise, Rafferty, Carilli, Sarazin and Blanton2005; Bîrzan et al. Reference Brzan, Rafferty, McNamara, Wise and Nulsen2004), or even become detectable as rejuvenated radio lobes (‘Phoenix’) following their adiabatic compression by cluster merger shocks (Enßlin & Gopal-Krishna Reference Enßlin and Gopal-Krishna2001). Furthermore, as pointed out in many studies, radio sources can significantly contribute to moderating or even preventing the cooling flows (Fabian Reference Fabian1994; Bîrzan et al. Reference Brzan, Rafferty, McNamara, Wise and Nulsen2004). Another dynamical effect, highlighted recently, is the bending of jets due to tension of the compressed ICM magnetic field (Chibueze et al. Reference Chibueze2021). The plethora of physical processes ongoing in galaxy clusters and the resulting rich phenomenology make them extremely interesting laboratories for studying the evolution of these largest gravitationally-bound structures in the universe. In this paper, we discuss new multi-band observations of a semi-relaxed cool-core cluster Abell 980 (A980), with particular attention to the ultra-steep-spectrum radio source complex embedded within the unusually hot ICM of this cluster, which display multiple indicators of episodic nuclear activity in its brightest cluster galaxy (BCG).
The BCG of the cluster A980 (RXC J1022.5+5006, z = 0.1582; Ebeling et al. Reference Ebeling1996) is known to possess a stellar halo of size
${\sim}80$
kpc, roughly co-spatial with the X-ray emission peak (see Figure 1). The cluster has an SZ mass of
$M_{500}^{SZ} = 4.73_{-0.32}^{+0.29} \times 10^{14} \mathrm{M_{\odot}}$
(Planck Collaboration et al. 2016) and a relatively high X-ray luminosity (
$L_X=7.1\times10^{44}$
erg s–1) originating from the ICM of size
${\sim}500$
kpc and a mean temperature
$T \sim7.1$
keV, Ebeling et al. Reference Ebeling1996). Its WENSS image at 330 MHz (FWHM =
$54{^{\prime\prime}}$
) gave the first hint of diffuse emission underlying a strong peak (Rudnick & Lemmerman Reference Rudnick and Lemmerman2009). From the radio observations with the GMRT (150/325 MHz) and VLA (1.5 GHz), the radio emission has been found to arise mainly from 3 components (A, B & C, Figure 1a), of which A and B are diffuse, with ultra-steep radio spectra (USS,
$\alpha<-2$
) and extending to the ICM outskirts, whereas the source C coinciding with the BCG is resolved into a smaller double radio source of size
$\sim$
55 kpc (Salunkhe et al. Reference Salunkhe, Paul, Gopal-Krishna and Bhagat2022; hereafter, S22). Here we present new observation-based evidence and arguments, particularly those emerging from the recently released high-sensitivity image with a 6 arcsec resolution at 144 MHz (LoTSS/DR2, Shimwell et al. Reference Shimwell2022), in order to reinforce the hypothesis associating all the 3 radio components (A, B, C) with the intermittent nuclear activity occurring in the BCG (S22). This 144 MHz image, when combined with the images obtained from the GMRT (325 MHz) and VLA (1.5 GHz) archival data, has enabled a clear delineation of the ultra-steep-spectrum (USS) components A & B and defining their structural and spectral details.
Parameters of the radio images.
Note: Resolutions: h - high, l - low; images are produced in w - WSCLEAN, s - SPAM and c - CASA. ‘*’ In WSClean, 5 arcsec corresponds to
${\sim}50$
k
$\lambda$
for this image.
Panel (a) Magenta contours of VLA L-band map at
$ 3, 9, 81\times\sigma=25\ \unicode{x03BC}$
Jy beam–1 (FWHM =
$9.0^{\prime\prime}\times8.7^{\prime\prime}$
) and blue contours (at
$ 3, 9, 27 \times \sigma=100\ \unicode{x03BC}$
Jy beam–1) of the GMRT-325 MHz map (FWHM =
$8.2{^{\prime\prime}}\times7.3{^{\prime\prime}}$
) are over-plotted on Pan-STARRS ‘i’ band optical image. Among the galaxies with known spectroscopic redshifts, the more likely cluster members (
$z=0.1582\pm0.0035$
; red squares) and the remaining galaxies (
$z<0.1547$
or
$z>0.1617$
; black squares) in the vicinity of the radio contours are shown. The southern object within the orange circle is discussed in Section 3.2. The south-west and north-east insets show the high-resolution VLA contours of C and B in magenta colour, respectively at
$[3, 6, 12, ....] \times \sigma$
and
$[3, 5, 7] \times \sigma$
(where,
$\sigma=30\ \unicode{x03BC}$
Jy beam–1 and FWHM=
$2.8^{\prime\prime}\times2.5^{\prime\prime}$
), superimposed on the same optical image. Panel (b) The grey colour Chandra X-ray map (with black contours at
$(2.68, 5.76, 8.80, 12.0, 15.0) \times 10^{-9}$
counts cm–2 s–1) is over-plotted with the LoTSS-2 144 MHz contours (magenta) at
$-3,5,10,20,...\times 150\ \unicode{x03BC}$
Jy beam–1 (FWHM=
$6^{\prime\prime}\times6^{\prime\prime}$
).
A
$\Lambda$
CDM cosmology is assumed with parameters
$H_{0}=70 {\mathrm{kms^{-1} Mpc^{-1}}}$
,
$\Omega_{0,m} =0.3$
,
$\Omega_{0,\Lambda}=0.7$
(Condon & Matthews Reference Condon and Matthews2018) throughout this paper. The physical scale for the images is 2.73 kpc arcsec–1.
2. The data and analysis
2.1 The radio data
A total of 314 min of on-source 325 MHz GMRT archival data (Project: ddtB020 & 17_073) and 56 min of EVLA L-band (1-2 GHz, B-array) archival data (Project: 15A-270 and 12A-019) were processed for this study. Data from RR and LL correlations of GMRT 325 MHz were reduced and imaged using the Source Peeling and Atmospheric Modeling (SPAM) pipeline (for details, see Intema et al. Reference Intema, Jagannathan, Mooley and Frail2017). The EVLA L-band data from RR and LL correlations were analysed using the CASA calibration pipeline and images were produced by performing several rounds of phase-only self-calibration and one amplitude and phase self-calibration. The images were made using the various weighting and uvtaper parameters listed in Table 1. The 144 MHz high-resolution (
$6^{\prime\prime}\times6^{\prime\prime}$
) map of LoFAR Two metre Sky Survey data release 2 (LoTSS-2) has been taken directly from the LoTSS-2 archive (Shimwell et al. Reference Shimwell2022).
The LOTSS-2 144 MHz and GMRT 325 MHz maps have been combined to produce the spectral index map. The SPAM calibrated GMRT visibilities at 325 MHz were taken to the WCSCLEAN algorithm (Offringa et al. Reference Offringa2014), to produce a 325 MHz image with the resolution of the LOTSS 144 MHz high-resolution map. Uniform weighting with a Gaussian taper of
$5^{\prime\prime}$
was used in WCSCLEAN. The small difference in the beam sizes of the two maps was then corrected by smoothing both maps to a common beam size of
$8^{\prime\prime}\times8^{\prime\prime}$
(circular), in the CASA task IMSMOOTH. The images were re-gridded to the same pixel grid using CASA IMREGRID task before smoothing. A minor astrometric offset between the LoTSS and GMRT images was corrected at the outset. Finally, the masking for these images was taken at
$3\sigma$
isophotes of 325 MHz map.
The spectral index was computed using the IMMATH task of CASA using the relation
\begin{equation}\alpha = \frac{\log(S_{\nu_2}/S_{\nu_1})}{\log(\nu_2/\nu_1)},\end{equation}
where
$S_{\nu_\#}$
and
$\nu_\#$
are the values of flux density and the frequency of observation, respectively.
The spectral index error has been calculated with the same IMMATH task using the relation given below (Kim & Trippe Reference Kim and Trippe2014)
\begin{equation}\alpha_{err}(\alpha_{\nu_2,\nu_1})=\frac{1}{\log(\nu_2/\nu_1)}\times\left[\frac{\sigma_{\nu_1}^2}{I_{\nu_1}^2}+\frac{\sigma_{\nu_2}^2}{I_{\nu_2}^2}\right]^{\frac{1}{2}}\end{equation}
with I as the total intensities at the respective frequencies at each pixel.
2.1.1 Computing magnetic field and the spectral age
The assumption that the energy density of cosmic rays and the magnetic fields are almost the same when the magnetised plasma in a synchrotron source is near its minimum energy state, allows one to estimate the strength of magnetic fields in such systems directly from their synchrotron radio emissions (Miley Reference Miley1980). Such estimate of magnetic field is often termed as the equipartition magnetic field and is calculated using the relation
\begin{equation} B_{eq} = 7.91 \left[ \frac{1+k}{(1+z)^{\alpha - 3}} \frac{S_0}{\nu_0^{\alpha} \theta_x \theta_y s} \frac{\nu_u^{\alpha+\frac{1}{2}} - \nu_l^{\alpha+\frac{1}{2}}}{\alpha+\frac{1}{2}}\right]^{\frac{2}{7}} \mathrm{\unicode{x03BC} G},\end{equation}
Here,
$S_0$
is the flux density (in mJy) at the observing frequency (
$\nu_0$
), k is the ratio of the energy content in relativistic protons to that in electrons,
$\nu_l$
and
$\nu_u$
are the lower and upper-frequency limits for computing the integrated radio luminosity, s is the path length through the source along the line of sight (in kpc), and
$\theta_x$
and
$\theta_y$
are source extents in two directions, measured in arcsec. For our case, we have adopted
$k=1$
,
$\nu_l =0.01$
GHz and
$\nu_u = 100$
GHz and
$\theta_x$
, the size and
$\theta_y$
, the path length have been assumed to be the same as the average size of the source.
To produce a spectral age map for the diffuse emission observed in A980, we fit JP model using the BRATS software package (Harwood et al. Reference Harwood, Hardcastle, Croston and Goodger2013). The maps used are at 144 MHz, 325 MHz and 1.5 GHz and the computation was done pixel-by-pixel in the region above 2
$\sigma$
of 1.5 GHz map. As input parameter, we use the injection index (
$\alpha_{inj}=0.7$
) computed using the BRATS software itself and an equipartition magnetic field (
$B_{eq}=2.6\ \unicode{x03BC}$
G) estimated using Equation (3) for the region A.
For the regions where data are available at only two frequencies, the spectral age has been calculated using the procedure described in Jamrozy et al. (Reference Jamrozy, Klein, Mack, Gregorini and Parma2004). In this model, with the assumption that a source has a uniform magnetic field, and ignoring any expansion losses, the age of a radio source with a spectral break frequency
$\nu_{br}$
is given by
\begin{equation} \tau = 1.59 \times 10^3 \nu_{br}^{-0.5} \frac{B_{eq}^{0.5}}{B_{eq}^2+B_{CMB}^2} \;\;\mathrm{Myr},\end{equation}
Here,
$B_{eq}$
is the equipartition field,
$B_{CMB} (\unicode{x03BC} G) = 3.18 (1+z)^2$
is the magnetic field equivalent to the cosmic microwave background (Longair Reference Longair2011) and
$\nu_{br}$
is the break frequency (in GHz). Obtaining precise break frequency from models would require observations at several well-spaced frequencies, in the absence of which we assume the break frequency as
$\nu_{br} = 144$
MHz, the lowest observing frequency available in our study.
2.2 The X-ray data
The level-1 event file of A980 from Chandra data archive (14 ks; ObsID 15105) was reprocessed following the standard data-reduction routine of CIAOFootnote a 4.11 and employing the latest calibration files CALDB 4.8.3. Events were screened for cosmic-rays using the ASCA grades, and were reprocessed by applying the most up-to-date corrections for the time-dependent gain change, charge transfer inefficiency, and degraded ACIS detector quantum efficiency. Periods of high background flares exceeding 20% of the mean background count rate were identified and removed using the lc_sigma_clip algorithm. This yielded 13.5 ks of net exposure time. The standard blank-sky data setsFootnote b were processed and re-projected to the corresponding sky positions and normalised to match the count rate in 10–12 keV energy range. Point sources across the ACIS field were identified using the CIAO tool wavdetect and removed. Thereafter, exposure correction was applied on the cleaned X-ray image (i.e., free from point sources and flares), using the mono-energetic exposure map created at 1 keV.
In addition, a 2D temperature map has been made using the contour-binning algorithm by Sanders (Reference Sanders2006). This algorithm generates a set of regions following the distribution of surface brightness such that each region has nearly the same signal-to-noise ratio. A total of 9 regions were thus delineated, each with a signal-to-noise ratio of 20 (
${\sim}400$
counts per bin), from the 0.5 to 3.0 keV image. We then extracted the X-ray spectra for these regions and fitted them with the model TBABS
$\times$
APEC using the method described in Section 2.2.1. Since the photon counts in individual patches are not statistically robust, we estimated the metallicity in circular annuli and used the values corresponding to different patches, for computing their temperatures. The estimated metallicity varies from the peak of 0.68
$\mathrm{Z}_\odot$
at the central annulus to 0.11
$\mathrm{Z}_\odot$
at the outermost annulus of the ICM.
The measured parameters of the radio components and of the integrated radio emission (see Figure 2b).
Panel (a) Spectral index distribution, derived by combining the GMRT (325 MHz) and LoTSS-2 (144 MHz) maps (see corresponding error map in 5). The colour map of spectral age produced by BRATS software (see Section 2.1.1) is shown within the inset (in units of Myr) and the corresponding error map in 6. Contours of the low-resolution map at 325 MHz are plotted at
$3, 12, 48, 192 \times 100\ \unicode{x03BC}$
Jy beam–1 over both the maps. Panel (b) Spectral plots for the integrated emission, and for the different components mentioned in the legend.
Panel (a): High-resolution (
$6^{\prime\prime}\times6^{\prime\prime}$
) map from LOTSS-DR2 (144 MHz) is shown in blue colour. The inset shows the corresponding contours for the central double radio source in yellow colour, at
$1, 2, 4, 8, ... \times 750\ \unicode{x03BC}$
Jy beam–1, over-plotted on the Pan-STARRS ‘i’ band optical image. Panel (b) GMRT map (325 MHz) is shown in blue colour and the inset shows the central double source in cyan colour contours, at
$1, 2, 4, 8, ... \times 650\ \unicode{x03BC}$
Jy beam–1
2.2.1 The global X-ray properties
The global properties of the cluster X-ray emission were determined by extracting a 0.5–8 keV spectrum of the X-ray photons from within a
$150^{\prime\prime}$
(408 kpc) circular region centred on the X-ray peak of the cluster. We excluded the central
$2^{\prime\prime}$
region, as well as the regions corresponding to the obvious point sources detected within the chip. A corresponding background spectrum was extracted from the normalised blank sky frame and appropriate responses were generated. The spectrum was binned such that every energy bin contains at least
$\sim$
20 counts and was then imported to XSPEC 12.9.1 for fitting, adopting the
$\chi^2$
statistics. The combined spectrum was fitted with an absorbed single temperature APEC model with the Galactic absorption fixed at
$N_{H}^{Gal} = 9.16\times10^{19}\mathrm{cm^{-2}}$
(Kalberla et al. Reference Kalberla, Burton, Hartmann, Arnal, Bajaja, Morras and Pöppel2005), letting the temperature, metallicity and normalisation parameter to vary. The best-fit resulted in the minimum
$\chi^2=219.01$
for 210 degrees of freedom.
3. Results and discussion
3.1 Multi-waveband view of the cluster A980
For a perspective of this system, we display in Figure 1a & b the multi-band overlays, followed by the derived radio spectral information in Figure 2a & b. Figure 1a shows the GMRT map (FWHM =
$8.2{^{\prime\prime}} \times 7.3{^{\prime\prime}}$
at 325 MHz) in which the components A, B and C are overlaid on an optical image of the field. The adjoining overlay displays the recently published LoTSS/DR2 map (FWHM =
$6^{\prime\prime}\times6^{\prime\prime}$
at 144 MHz) superposed on the Chandra X-ray image (Figure 1b). In Figure 2a, we present the spectral index map derived by combining the LoTSS/DR2 and GMRT maps, after matching them in angular resolution. The brighter components A and C, which are barely separated in the TGSS-ADR1 map with a 25 arcsec beam at 150 MHz (Intema et al. Reference Intema, Jagannathan, Mooley and Frail2017), can now be clearly demarcated in the LoTSS 144 MHz map, enabling a reliable extension of their spectra down to this low-frequency (Figure 2b; Table 2). Figure 3 shows these two radio maps in blue-colour. Lastly, Figure 4 displays the LoTSS-2 high-resolution contour map overlaid on the ICM temperature map in which a hot patch (
$T \sim 11$
keV) is present just to the north-east of the cool core of this cluster.
Chandra X-ray temperature (keV) map overlaid with the radio contours (at
$5,20,80,320 \times 150\ \unicode{x03BC}$
Jy beam–1) of the LOTSS-2 map.
3.2 Parent galaxy of the USS radio components A and B
The radio ridge of component A, extending towards B (Figure 1a & b) strongly suggests that these two USS components are physically linked relic lobes (
$\alpha<-2.0$
; Figure 2b). The lack of a terminal hot spot means that the radio ridge seen in A is unlikely to be a manifestation of back-flowing synchrotron plasma. This is clearly borne out by the radio spectral gradient derived here, which shows the spectrum to become flatter going inward (Figure 2a). This provides a vital observational evidence for the hypothesis that the two lobes are aged synchrotron plasmons produced by one galaxy during its active phase and the oldest synchrotron plasma, propelled by buoyancy in the cluster potential, has now reached the ICM outskits (S22). The spectral gradient derived here shows the rising USS plasmons being trailed by younger plasma. Observational clue in support of this buoyancy interpretation (Gull & Northover Reference Gull and Northover1973; Sakelliou, Merrifield, & McHardy Reference Sakelliou, Merrifield and McHardy1996) comes from the good alignment between the front edge of each component with the X-rays contour at that location (Figure 1b). The observed lateral spreading of the lobes is expected to set in when a plasmon has buoyantly risen to the iso-entropy surface within the ICM (e.g., Kaiser Reference Kaiser2003). Secondly, the spectral ageing analysis (Section 2.1.1) gives an age of
${\sim}260$
Myr for the outer parts of the relic lobes A and B, which can easily allow the proposed long buoyant drift through the ICM with an expected average speed roughly a third of the ICM sound speed (Churazov et al. Reference Churazov, Brüggen, Kaiser, Böhringer and Forman2001), c
$_s$
(ICM), which is
${\sim}1\,350$
km s–1 for
$T_{ICM} \sim 7$
keV (Section 1). The alternative scenario that A and B are aged radio trails of a Wide-Angle-Tail (WAT) source is less appealing because the radio contours of both A and B in the more sensitive LoTSS map (Figure 1b) show no sign of trailing off towards the expected north-eastern direction. Here it may also be mentioned that the southward radio spur extending from component B is a feature unrelated to B since it is contributed by two point-like radio sources detected in the 1.5 GHz EVLA map, which coincide with the two galaxies shown in black squares in Figure 1a (see, also the north-eastern inset).
In contrast to FR II sources, the spectrum is seen to become flatter going inwards along the radio lobe A (Figure 2a). However, such a spectral gradient is seen in the cluster radio galaxy 3C 388 and explained in terms of intermittent jet activity in its host galaxy (Roettiger et al. Reference Roettiger, Burns, Clarke and Christiansen1994; Brienza et al. Reference Brienza2020). Conceivably, the inner part of the radio ridge in component A is a manifestation of younger synchrotron plasma injected by residual low-level jet activity of the host galaxy (see, e.g., Sabater et al. Reference Sabater2019), while the wider outer part of A (and B) are the remnants from the earlier, main active phase of the host galaxy. The spectral ageing analysis using BRATS software gives
${\sim}90$
Myr for the age of the synchrotron plasma near the centre, gradually rising to
$>240$
Myr for the outer parts of lobe A (see inset of Figure 2a). The regions of lobe B and the part of A, for which radio emission at only two frequencies are detected, the ages have been computed following the procedure described in (Jamrozy et al. Reference Jamrozy, Klein, Mack, Gregorini and Parma2004) and adopting the JP model (Section 2.1.1), and are found to be
$255\pm45$
and
$261\pm32$
Myr, respectively. Interestingly, Figure 1a does not show even a moderately bright optical galaxy near the central location between A and B. Of the two optical objects seen within the orange circle, the northern one is classified as a star and the lower one as a galaxy (SDSS/DR7, Abazajian et al. Reference Abazajian2009), lacking radio detection (Figure 1a). The values of apparent magnitude (m
$_r$
) of this galaxy are given as 19.96
$\pm$
0.05 (PANSTRASS-1, Chambers et al. Reference Chambers2016) and 20.64
$\pm$
0.09 (SDSS/DR7, Abazajian et al. Reference Abazajian2009). Its redshift is not known and it would be desirable to measure it, but assuming that it does lie within the cluster A980, the absolute magnitudes would be M
$_r$
= –19.44
$\pm$
0.07 and –18.76
$\pm$
0.10, respectively. These can be compared with M
$_r$
= –22.8
$\pm$
0.5 for the hosts of radio galaxies at
$z < 0.5$
, for the same cosmology as the one adopted in the present work (Sbarufatti, Treves, & Falomo Reference Sbarufatti, Treves and Falomo2005). Thus, a comparison with the above PANSTRASS-1 and SDSS/DR7 based estimates of M
$_r$
implies that the galaxy within the orange circle, if indeed a member of A980, is under-luminous by 3.36-mag and 4.04-mag, respectively, that is, 6.6
$\sigma$
and 7.9
$\sigma$
, where
$\sigma$
is the combined rms error. This makes it a highly improbable host of the relic lobe pair A - B. Discarding this unlikely host, the plausibility favours the scenario (Section 1) that the relic lobes A and B were both created by the BCG during its previous major episode of jet activity when it was near the central location between A and B. Since then, the galaxy has drifted towards the cluster centre (BCG in Figure 1a) and entered a new phase of intense jet activity which is identified with the creation of the smaller double radio source with an active central core (L1, L2 in the inset of Figure 1a). Thus, in this scenario, the buoyant rise of the diffuse relic lobes A and B within the ICM, following the near-cessation of intense jet activity in their parent galaxy (BCG), and the migration of this galaxy towards the cluster’s gravitational centre are two concurrent processes. Besides the obvious fact that the BCG is massive enough to engender a powerful large double radio source, this scenario is supported by the following additional arguments:
(a) The radial migration of the active galaxy (BCG) to the cluster centre, as required in the present scenario, is the natural trajectory for the galaxy to take (a requirement of migration in the opposite direction would have rendered the scenario untenable). Note also that only a modest average speed of
${\sim}725$
km s–1 is needed for the galaxy to traverse the (projected) radial distance of
${\sim}70$
kpc from the A–B axis to its present location near the cluster centre, during the estimated spectral age (
${\sim}95$
Myr) of the plasma around the mid-point between the relic lobes A and B (Section 3.1; Figure 2a). Such a speed would be entirely consistent with the observed velocity dispersion of 1 033 km s–1 for the galaxies in this cluster (Rines et al. Reference Rines, Geller, Diaferio and Kurtz2013).
(b) An important piece of supporting evidence contributed by the LoTSS map (Figure 3a) is the faint radio spur emerging orthogonally from the A–B axis and connecting to the BCG, thus tracing the proposed radial trajectory for the drift of the BCG into the cluster centre. Although a hint of this spur is visible in the GMRT map (Figure 3b), its independent detection in the LoTSS map is crucial, for this confirms the reality of this weak feature and thus provides an important clue. Quite plausibly, this radio spur arises from some of the magnetised relativistic plasma of the lobe A being dragged along in the wake of the moving BCG, with additional contribution coming from the residual low-level activity in the moving BCG (see, e.g., Sabater et al. Reference Sabater2019).
(c) The A–B axis is parallel to the axis of the young double (L1, L2). This accords with the frequently observed near-constancy of the jet direction between successive episodes of jet activity in DDRGs (e.g., Saikia & Jamrozy Reference Saikia and Jamrozy2009). A similar case for near-constancy of the jet direction has been made for X-shaped radio galaxies, as well (Cotton et al. Reference Cotton2020 and references therein).
(d) The hypothesis that the USS sources A and B have a common host galaxy, is also supported by the presence of the extra-hot patches of the ICM, seen between A and B (shown in red/orange colour in Figure 4). The ICM temperature at these patches is
${\sim}10\%$
higher than the surroundings. Such a temperature enhancement associated with powerful jet activity in clusters has been observed in numerical simulations (Raouf et al. Reference Raouf, Shabala, Croton, Khosroshahi and Bernyk2017). Since both hot patches fall outside the cool core, their high temperature is expected to have persisted over the age of the relic lobes A and B.
The recurrence of jet activity in the BCG, which is germane to the present scenario, is also entirely consistent with the notion that cool cores of clusters are propitious sites for triggering jet activity (e.g., Burns Reference Burns1990; Fabian et al. Reference Fabian, Sanders, Taylor, Allen, Crawford, Johnstone and Iwasawa2006; Mittal et al. Reference Mittal, Hudson, Reiprich and Clarke2009). A plausible reason is that warm extended filaments and cold clouds condensing out of the ICM would accrete steadily onto the SMBH in the cluster central galaxy (e.g., Pizzolato & Soker Reference Pizzolato and Soker2005; Sharma et al. Reference Sharma, McCourt, Quataert and Parrish2012; Voit & Donahue Reference Voit and Donahue2015). Moreover, their accretion rate can be boosted by almost two orders of magnitude over the Bondi prediction, as inelastic collisions between the gas clouds in the central region cause angular momentum cancellation, triggering a fresh episode of jet activity in the dominant central galaxy (Gaspari, Ruszkowski, & Oh Reference Gaspari, Ruszkowski and Oh2013).
Purely from an observational standpoint, the most dramatic manifestations of recurring jet activity are the so-called, ‘double-double’ radio galaxies (DDRGs) in which the host galaxy is straddled by an inner (younger) and an outer (older) pair of radio lobes (e.g., Schoenmakers et al. Reference Schoenmakers, de Bruyn, Röttgering and van der Laan2000; Saikia & Jamrozy Reference Saikia and Jamrozy2009; Morganti Reference Morganti2017). Still, rarer manifestations are the so-called ‘triple-double radio sources’ (Brocksopp et al. Reference Brocksopp, Kaiser, Schoenmakers and de Bruyn2007) and the unique case where such a morphology is associated with the galaxy J140948.85-030232.5, a spiral within a galaxy cluster (‘SPECA’, Hota et al. Reference Hota2011). The novel aspect of the source in A980 is the large positional shift of the parent galaxy between its major episodes of jet activity. Consequently, the two laterally well-separated pairs of old (A, B) and young (L1, L2) radio lobes, whose axes are parallel but which would customarily be deemed as two unrelated double radio sources, are actually the ‘outer’ and ‘inner’ lobe pairs of a single radio galaxy, which have lost their usual collinearity and got detached from each other owing to a lateral motion of the parent galaxy. Very likely, in this case of a detached DDRG (dDDRG), both the large lateral displacement of the parent galaxy between its activity episodes and the prolonged preservation of the identity and detectability of the older lobe pair, have been facilitated by their location inside a relaxed cluster.
Although we have argued that the radio components A and B are physically associated as a pair of fossil radio lobes of a single parent galaxy, a potential alternative scenario would be that A and B are independent radio sources. Under this hypothesis, the parent galaxy of B might be one of the two ellipticals coinciding with its southward radio spur, each of which has a distinct radio counterpart detected in the high resolution EVLA map, as seen within the north-eastern inset to Figure 1a. At the same time, the component A may be associated with the triple radio source C, together forming a Narrow-Angle-Tail (NAT) radio source hosted by the BCG. Although, this hypothesis cannot be ruled out, it seems improbable in view of the following radio indicators:
(i) The radio component A shows a prominent elongation towards B and, moreover, the two have similarly ultra-steep radio spectra (Figure 2a) and diffuse radio morphology (Figures 1a and 3b). All this suggests a common origin of A and B.
(ii) Far from being continuous, the radio ‘connection’ between A and C is very faint, almost like a discontinuity in surface brightness (Figure 3b).
(iii) Instead of an expected steady spectral steepening from C towards A, a significant spectral index discontinuity is observed between them. The faint radio spur linking A and C has
$\alpha=-1.85\pm0.18$
and is flanked by regions of flatter spectra (
$\alpha=-1.00 \pm 0.14$
to the north and
$-0.97\pm0.05$
to the south, see Figure 2a). It is conceivable that the flatter spectrum region to the north is due to a better confinement of the radio plasma due to the very high temperature of the ICM near that location (Figure 4). However, an expected surface brightness enhancement due to the better confinement is not observed.
(iv) In the moving BCG scenario (also essential for the NAT interpretation), radial infall is the natural trajectory for the BCG to take for arriving at its present location near the cluster centre. Any radio trail of the BCG would hence be expected to be elongated in the radial direction, whereas the component A extends almost orthogonally to that direction (Figure 3b).
3.3 The other known dDDRG candidates
3C 338: This
${\sim}80$
kpc long radio source, hosted by the multi-peaked cD galaxy in Abell 2199, consists of (i) a small double radio source having a twin-jet emanating from a radio core, and (ii) an offset pair of much larger diffuse USS lobes joined by a collimated jet-like radio feature with
$\alpha \simeq -1.9$
(Burns, Schwendeman, & White 1983; Ge & Owen Reference Ge and Owen1993). These authors have considered the possibility of this USS lobe pair being a relic from past activity in the galaxy which has since moved laterally northward and produced the younger lobe pair. However, they have also expressed caution about this scenario on account of the high brightness and strong collimation of the USS radio feature, the putative ‘relic jet’. In fact, this feature bears a striking resemblance to the recently discovered ‘collimated synchrotron threads’ in the WAT radio source ESO 137-006 (Ramatsoku et al. Reference Ramatsoku2020).
4C 35.06 (B2 0258+35B): The complex radio morphology of this source in the cluster Abell 407 (z = 0.047) has been unravelled through the radio imaging with VLA (1.4/4.9 GHz)/LOFAR (62 MHz) (Shulevski et al. Reference Shulevski2015) and GMRT (150/235/610 MHz) (Biju et al. Reference Biju, Pandey-Pommier, Sunilkumar, Dhurde, Bagchi, Ishwara-Chandra and Jacob2014, Reference Biju2017). The cluster core hosts a uniquely tight grouping of nine ellipticals (‘Zwicky’s Nonet’) confined within just
${\sim}50\,$
kpc. Two of those ellipticals are identifiable with two double radio sources of very unequal extents, whose axes are parallel but offset by
${\sim}11$
kpc (Biju et al. Reference Biju2017). The larger radio source comprises a 400 kpc long pair of helical jets/lobes whose spectral index (
$\alpha$
) gets as steep as -2 in the outer parts. Although its proposed host galaxy is 3–4 mag fainter than typical hosts of large radio galaxies, this mass deficit has been attributed to a severe tidal stripping of its stellar envelope due to the uniquely dense galaxy environment of the Nonet in which this galaxy is located (Biju et al. Reference Biju, Pandey-Pommier, Sunilkumar, Dhurde, Bagchi, Ishwara-Chandra and Jacob2014, Reference Biju2017; see, also Graham & Scott Reference Graham and Scott2015). In contrast, Shulevski et al. (Reference Shulevski2015) have asserted that both double radio sources have been created by the same luminous elliptical UGC 2489 which now exhibits a VLBI radio core and is seen to coincide with the smaller double source. In their scheme, the larger radio source got created
${\sim}35\,$
Myr ago in a previous active phase of this massive elliptical, which has since moved laterally by
${\sim}11$
kpc to its present location and entered a new active phase, resulting in the smaller double radio source. However, this interpretation remains uncertain due to the afore-mentioned debate over the host galaxy of the larger double radio source. Furthermore, the nature of the smaller ‘double radio source’ is itself somewhat unclear, since each of its radio ‘lobes’ coincides with an elliptical galaxy belonging to the Nonet (Figure 1 of Shulevski et al. Reference Shulevski2015; Figure 3 of Biju et al. Reference Biju2017).
4. Conclusions
Based on multiple observation-based evidences, we have argued that the complex radio emission in cluster Abell 980 arises mainly from two double radio sources created during two episodes of activity in the BCG, which has been drifting towards the cluster centre. Both lobes of the older radio source, spanning
${\sim}350$
kpc, are now diffuse. Their ultra-steep spectrum (
$\alpha < -2$
) and flattened appearance near the extremities are consistent with the notion that these two relic lobes have risen buoyantly towards the cluster periphery during their estimated spectral age of
${\sim}260$
Myr. The jet activity leading to the creation of these relic twin-lobes is probably also responsible for the two patches of enhanced ICM temperature observed between the lobes, consistent with recent numerical simulations. The axis of the younger double radio source coincident with the BCG is found to be parallel to the axis of the relic lobe pair, albeit at a lateral offset by
${\sim}70$
kpc. It is argued that the old (relic) lobe pair and the younger double radio sources are probably not independent sources, but, represent a double-double radio galaxy whose old and young lobe pairs have got laterally ‘detached’ due to the drift of their (common) parent galaxy towards the cluster centre. It is further noted that the two previously reported candidates for such detached Double-Double Radio Galaxy (dDDRG) seem to be less viable examples of this phenomenon.
Acknowledgement
G-K acknowledges a Senior Scientist fellowship from the Indian National Science Academy. SP would like to thank INSPIRE Faculty Scheme of DST, Govt. of India (code: IF-12/PH-44) for the research fund. S. Salunkhe would like to thank ‘Bhartratna JRD Tata Gunwant Sanshodhak Shishyavruti Yojna’ for a doctoral fellowship. S. Sonkamble acknowledges the financial contribution from ASI-INAF n.2017-14-H.0 (PI A. Moretti). We thank Prof. Paul Wiita for his constructive comments on the manuscript. We thank the staff of the GMRT that made these observations possible. GMRT is run by the National Centre for Radio Astrophysics of the Tata Institute of Fundamental Research.
