Ultra-steep-spectrum Radio “Jellyfish” Uncovered in A2877

Synchrotron sources typically exhibit power-law behavior in their spectra, such that the observed flux S is related to frequency ν by the relation S ∝ ν^α . The spectral index, α, is typically around −0.7 for the lobes of active galactic nuclei (AGNs), and other mechanisms, such as radio halos, relics, and AGN remnants, have been observed to have spectra as steep as −2. For most of these sources, the spectral index is a well-understood parameter that results from the original injection energy and the aging dynamics of the system.

With the advent of new low-frequency telescopes, in particular the Murchison Widefield Array (MWA; Tingay et al. 2013) and the LOw Frequency ARay (LOFAR; van Haarlem et al. 2013), it was widely expected that we would uncover a large, hitherto undetected population of ultra-steep-spectrum (USS) sources only detectable at low frequency (e.g., Enßlin & Röttgering 2002; Cassano et al. 2012, 2015; van Weeren et al. 2019). It was believed that traditionally higher-frequency radio observations introduced an observational bias against sources whose spectra rapidly declined in luminosity with frequency. But these expectations have not been realized. Instead, it appears that the conditions required to produce low-frequency USS synchrotron emission are not particularly common. And for those that we have found, it is all the more important for us to understand the unique conditions that make them possible.

To date, the steepest reported source is in A1033, the so-called “Gently ReEnergised Tail” (GReET; de Gasperin et al. 2017), with an integrated spectral index α = −3.86(3). The GReET is composed of ancient plasma originally ejected from a wide-angle tail radio galaxy and since reaccelerated. The reacceleration mechanism remains unknown, but the authors considered two scenarios: adiabatic compression due to weak shocks and stochastic reacceleration driven by complex turbulence in the tail and interaction with the surrounding intracluster medium (ICM). In this particular instance, they concluded that the latter scenario was more likely. The spectral steepness of the GReET was so steep that it was only visible at low frequency and became undetectable above 323 MHz.

The GReET is a kind of radio phoenix, which is a class of synchrotron sources that arises from the reacceleration of ancient but still mildly relativistic (γ > 100) “fossil” electron populations, usually old AGN cocoons or remnants. Other examples include phoenix in A1664 (e.g., Kale & Dwarakanath 2012), A2256 (van Weeren et al. 2009), and, recently, A1914 (Mandal et al. 2019). In radio phoenix, the underlying fossil electron population is not well mixed with the ICM, and their morphology usually traces out the underlying AGN lobes or tails, albeit made more complex due to diffusion, buoyancy effects, turbulence, and the reacceleration mechanism itself. Radio phoenix rely on a mechanism to reignite an otherwise aged and faded AGN remnant, and adiabatic compression has been suggested as one of these mechanisms (Enßlin & Gopal-Krishna 2001), possibly caused by cluster–cluster interaction. In most radio phoenix to date, however, there is no evidence of shocks, and some are even found in relaxed cool-core clusters, implying that major merger events are not required as a trigger. Other mechanisms have been suggested, such as the proposed cool-core “sloshing” in the Ophiuchus cluster to account for its giant radio fossil (Giacintucci et al. 2020). Radio phoenix are typically found close to the central cluster region and are usually a few hundred kiloparsecs in size (Feretti et al. 2012). They often exhibit USS (α < −1.5) and display spectral curvature, and their spectral index maps typically do not indicate large-scale coherence or trends (see van Weeren et al. 2019 and references therein).

Here we report on the discovery of a diffuse USS radio source to the northwest of A2877, also likely a radio phoenix. The source spans ∼740″ in width, has two bright peaks of emission associated with cluster members, and has tentacles of emission that extend south toward the cluster core, giving the impression of a jellyfish. The “USS Jellyfish” was first detected in images from the Galactic and Extra-galactic All-sky MWA survey (GLEAM; Wayth et al. 2015) by searching for steep-spectrum sources. Due to its extreme spectral steepness, it had no detectable counterpart in the highest-frequency GLEAM images centered at 200.5 MHz and thus did not form part of the original GLEAM catalog (Hurley-Walker et al. 2017). GLEAM was conducted with the lower-resolution phase I of the MWA (Tingay et al. 2013), resulting in a blended source that made determining its morphology difficult. This prompted the follow-up radio observations we present here with the higher-resolution MWA phase II (extended configuration; Wayth et al. 2018) and the Australia Telescope Compact Array (ATCA; Frater et al. 1992), as well as reprocessed archival XMM-Newton X-ray observations.

Throughout, we assume a flat ΛCDM cosmology with Hubble constant h = 0.677 and matter density Ω_m = 0.307, of which the the baryonic density is Ω_b = 0.0486. All coordinates are with respect to the J2000 epoch. All stated errors indicate one standard deviation.

1.1. A2877

2.1. ATCA

2.2. MWA

2.3. Spectral Index and Image Map

2.4. XMM-Newton

We present the USS Jellyfish in Figure 1 as a composite image combining the optical Digitized Sky Survey (background) with wavelet-cleaned XMM X-ray data (magenta overlay) and MWA 118.5 MHz radio data (red-yellow overlay). Blue circles indicate known radio-emitting cluster sources from Hopkins et al. (2000), while blue diamonds indicate probable cluster members based on available redshift data. The X-ray emission shows the core overdensity of A2877, to which the radio emission is offset to the northwest. The radio emission resembles a jellyfish, with western and eastern peaks of emission in the head, and likewise western and eastern tentacles that descend south toward the cluster core. The full angular extent of the USS Jellyfish from east to west is ∼740″, which at the redshift of A2877 corresponds to ∼370 kpc in projection.

Zoom In Zoom Out Reset image size

Figure 2 shows the emergence of the USS Jellyfish from near undetectability at 215.5 MHz to an integrated flux of 1.10(15) Jy at 87.5 MHz, approximately 275 times more luminous. At this lowest frequency, the integrated flux corresponds to a total radio luminosity of L_87.5MHz = (1.59 ± 0.22) × 10²⁴ W Hz⁻¹, assuming a redshift z = 0.0238.

Zoom In Zoom Out Reset image size

The steepness of the spectra is extreme and shows significant curvature, as can be seen in Figure 3. This figure shows the integrated flux measured in each band, linearly interpolated (gray dashed line), while Table 3 additionally provides pairwise spectral indices. If we attempt a power-law fit (S ∝ ν^α ) to the three central bands, we find a spectral index value of $\alpha =-{5.97}_{-0.48}^{+0.40}$.

Zoom In Zoom Out Reset image size

Table 3. The Integrated Flux of the USS Jellyfish at Each MWA Band and Pairwise Spectral Indices

Frequency	Flux Density	Spectral Index
MHz	Jy
87.5	1.10 ± 0.06	${\alpha }_{87.5}^{118.5}=-{2.16}_{-0.27}^{+0.27}$
118.5	0.57 ± 0.03	${\alpha }_{118.5}^{154.5}=-{4.87}_{-0.39}^{+0.38}$
154.5	0.16 ± 0.01	${\alpha }_{154.5}^{185.5}=-{7.8}_{-1.3}^{+1.1}$
185.5	0.038 ± 0.008	${\alpha }_{185.5}^{215.5}\lt -9.2$
215.5	0.003 ± 0.007

Note. We provide only an upper bound for the highest-frequency pairwise spectral index, as the integrated flux measurement at 215.5 MHz is consistent with zero.

Download table as:  ASCIITypeset image

Unsurprisingly, higher-frequency observations show no extended emission that is coincident with the USS Jellyfish. The previously stated spectral index would give a total integrated flux at 887 MHz of just 4 μJy. Thus, at 887 MHz, the respective RACS mosaic, with a local noise of 220 μJy beam⁻¹, shows no emission besides point-source emission from A2877 galaxy members; additional convolution steps to emphasize extended diffuse emission also do not reveal any signal. Similarly, the lowest-frequency band of our ATCA observation at 1548.5 MHz, with a local noise of 52 μJy beam⁻¹, does not reveal any of the extended emission of the Jellyfish, nor do the higher-frequency ATCA bands.

There are three radio-emitting members of A2877 entangled in the USS Jellyfish that are detectable in the 215.5 MHz MWA image, and there is a fourth background source. In Figure 4, we show the combined ATCA observation centered at 2223.5 MHz overlaid with contours from the MWA at 215.5 and 118.5 MHz. In the highest-frequency MWA image, we identify two sources of emission, labeled A and B, that at lower frequencies are closely associated with the bright western and eastern peaks at the head of the USS Jellyfish. At 215.5 MHz, the radio emission at A and the E/S0 galaxy ESO 243 G 045 are well aligned. There is evidence of a slight southeast elongation of the emission, and indeed, at lower frequencies, the western peak shifts slightly southeast from A. Source B is offset slightly to the southwest of S0 galaxy WISEA J010946.55–454657.4. The northwest alignment of the emission of source B at 215.5 MHz remains visible at lower frequencies, suggesting continuity of the emission source. Hopkins et al. (2000) had previously classified both of these radio sources: ESO 243 G 045 as a low-luminosity AGN and WISEA J010946.55–454657.4 unambiguously as a Seyfert 2 galaxy. Source cD is the central cluster galaxy, IC 1633, and in the lower-frequency bands, the western tentacle bridges this source to the rest of the emission. It is identified by Hopkins et al. (2000) as hosting an AGN, and both their ATCA observation and ours show nearby faint emission to the northeast that may indicate a small jet. A background radio galaxy (z = 0.545; Afonso et al. 2005) with resolved ∼15″ jets is also visible in the MWA images and is labeled E.

Zoom In Zoom Out Reset image size

The spectral index map (Figure 5) shows no overall trend in the spectral index values across the full extent of the emission. Instead, we observe a patchwork of islands of shallower emission, surrounded by more diffuse and steeper emission. Both the western and eastern peaks of emission at the head of the USS Jellyfish are associated with flatter spectra, as is the emission coincident with the cD galaxy itself and the background radio galaxy E. Note that the lower spectral index values at the western and eastern peaks are not simply caused by a bias introduced from point sources A and B, since both are offset to their respective galaxy counterparts. The rest of the extended emission has a spectral index of around −4, and at some edges and along the western tentacle, the spectrum tends steeper still.

Zoom In Zoom Out Reset image size

From the X-ray, we observe that A2877 strongly resembles a cool-core cluster, with the core being X-ray bright and having a temperature of ∼2 keV. In Figure 6, we show the exposure-corrected and background-subtracted X-ray surface brightness profile of the northwest sector in the 0.5–2.5 keV band (blue points) alongside the radio surface brightness profile of the USS Jellyfish at 118.5 MHz (dashed gray line). There is no evidence for a shock or cold front in this sector; however, it remains possible that this is due to the poor data in this sector or that the feature is intrinsically faint due to projection effects.

Zoom In Zoom Out Reset image size

The USS emission on cluster-sized scales points to only a handful of plausible mechanisms to account for its spectral steepness: radio relics and halos, AGN remnants, and reaccelerated AGN plasma or “phoenix.” ¹¹ We can readily discount halos, as the emission is significantly offset from the cluster core, and radio halos trace the baryonic content of the cluster. Radio relics trace large-scale cluster shocks and usually appear toward the periphery of a cluster, although they have been observed more centrally. Relics have values for α of typically −1 to −1.5 and often display large-scale coherency in their spectral index map, indicating the direction of the shock, which we do not observe here. Their morphology is usually long and narrow, as they trace the bow of the shock. The spectral index of a relic is related to the shock strength and, assuming diffusive shock acceleration (DSA) on a purely thermal plasma and a conservatively shallow value of α_inj < −3, we find a shock Mach number ${ \mathcal M }\lt 1.3$. ¹² Modeling of DSA by Hoeft & Brüggen (2007) shows that the efficiency of electron acceleration is strongly dependent on the Mach number, and for shocks ${ \mathcal M }\lt 3$, their models show a rapid and exponential decrease in this efficiency. In the case of the USS Jellyfish, and without recourse to a significant fossil electron population (e.g., see models by Kang & Ryu 2011; Pinzke et al. 2013), such a truly gentle shock is an exceptionally inefficient electron accelerator. For these reasons, we deem the radio relic hypothesis unlikely.

An AGN remnant is another potential scenario, as models for synchrotron aging predict a steepening spectral index, increasing spectral curvature, and decreasing “break frequency” (e.g., Kardashev 1962; Pacholczyk 1970, “KP model”; Jaffe & Perola 1973, “JP model”). None have been observed to be as steep as observed here; however, it is possible that we are observing the remnant in a frequency range above the break frequency. If we assume the break frequency to be <70 MHz and the cluster magnetic field strength to be in the range of ≈0.01–1 μG, ¹³ we find a synchrotron age for the USS Jellyfish of at least 24–220 or 53–490 Myr, based on the KP and JP models, respectively. This is significantly older than any known AGN remnant, and indeed, recent simulations suggest rapid cooling and dimming of radio lobes after AGN shutoff due to adiabatic losses on the order of just millions of years (Godfrey et al. 2017; English et al. 2019). It is unlikely that the USS Jellyfish is an undisturbed AGN remnant.

Instead, our principal hypothesis is that it is far more likely that the USS Jellyfish is composed of multiple radio phoenix—a polyphoenix—triggered by a common large-scale shock or compression occurring to the northwest of A2877. The strongest evidence for this consists of the western and eastern peaks of emission and their association with nearby point sources A and B, respectively. Both peaks are the least steep components of the spectral image map, indicating a cocoon of younger, more energetic electrons that envelop cluster members ESO 243 G 045 and WISEA J010946.55–454657.4. We suggest that the shallower-spectrum component at the western peak, though it is offset to the southeast of source A, is directly related to emission from ESO 243 G 045. We also suggest that the elongated eastern peak is the product of a pair of weak AGN jets originating from source B, WISEA J010946.55–454657.4; the slight offset likely indicates that this activity was historic. Moreover, the eastern spur of the jellyfish head can be explained by deflection of the southeastern jet of source B by the denser, more central ICM. While both electron populations would have rapidly dimmed if unperturbed, we suggest that they have been compressed and reignited by some external mechanism(s).

The large-scale, more diffuse emission of the USS Jellyfish, as well as its tentacles, remains difficult to explain. The eastern extension from ESO 243 G 045 that is visible at 185.5 MHz is suggestive of either a lobe from a previously active epoch of AGN activity or a tail. If it is the latter, one plausible explanation is that both tentacles are the reenergized tail of ESO 243 G 045 that sweeps south toward the core of the cluster. An alternative explanation would be to invoke the cD galaxy IC 1633 as a third electron source, since the western tentacle, and to some degree the eastern, establishes a bridge of emission to this source.

The fact that multiple disparate electron populations have been reaccelerated strongly suggests a common large-scale reacceleration mechanism. As the XMM observation shows, there is presently no evidence of a shock in the northwest sector; moreover, it appears that A2877 is a relaxed, cool-core cluster and therefore unlikely to be currently subject to disruptive merger events or other large-scale structure formation processes that could trigger such a large-scale shock. Cool-core sloshing, which is identified in the X-ray by the presence of spiral or arc-shaped cold fronts about the central core, is an alternative mechanism that is triggered by instabilities acting upon the deep gravitational well at the centers of cool-core clusters (e.g., see Ascasibar & Markevitch 2006; Ghizzardi et al. 2010; Vazza et al. 2012). Indeed, the previously identified northern substructure in A2877 may provide just such an instability. While this mechanism has typically been invoked to explain centrally located “mini-halos” (e.g., Giacintucci et al. 2014), cold fronts associated with cool-core sloshing have been observed up to 1 Mpc away (Simionescu et al. 2012; Rossetti et al. 2013). However, until we can obtain both high-resolution and high-sensitivity X-ray observations of the cluster, the existence and nature of any shocks in the system must remain purely speculative.

Of special note, the spectral index map of the USS Jellyfish indicates spectra that are still strongly correlated with the original plasma age. In a standard DSA scenario, a strong shock would have imprinted the strength of the shock itself on the spectrum of the plasma and would display the same kind of large-scale coherence observed in the spectral index maps of many relics. We can thus infer a particularly gentle shock, and, based on our current understanding of DSA physics and the inefficiency of weak shocks, we can also infer the existence of a significant population of suprathermal electrons throughout the northwest sector of A2877, further reinforcing the phoenix hypothesis. Additionally, adiabatic compression of AGN cocoons akin to that originally described in Enßlin & Gopal-Krishna (2001) is another complimentary mechanism. Indeed, due to the higher speed of sound in the cocoon environment, shock waves will only poorly penetrate the AGN cocoons, and DSA would thus be of negligible effect. Their modeling suggests that adiabatic compression can boost the luminosity of the AGN cocoons without a significant flattening of the original spectrum, thus preserving the underlying aged spectra. Such compression, however, is unlikely to explain the large-scale diffuse radio emission observed exterior to the AGN cocoons. We also raise the possibility of a third reacceleration mechanism, local turbulence in the wake of a weak shock between interacting AGN lobes, powering Fermi II acceleration processes.

To explore a possible formation scenario for the USS Jellyfish, we turn to a recent suite of magnetohydrodynamical simulations by Vazza et al. (submitted). In one of these scenarios, we have identified a polyphoenix akin to the USS Jellyfish as a transient feature in the evolving ICM. In these simulations, light jets (i.e., filled by hot gas and magnetic fields) were released by AGN particles in the simulation within a forming galaxy cluster with a total mass of M₁₀₀ ≈ 1.5 × 10¹⁴ M_⊙. For the runs used in this comparison, the AGN jets were initiated at the same epoch for the four most massive AGNs in the simulation, all located within ∼Mpc³ (comoving) and each releasing ∼10⁵⁷ erg of feedback energy into the surrounding medium in the form of kinetic, thermal, and magnetic energy. The model for the spectral energy evolution of relativistic electrons is similar to the one presented in Vazza et al. (2021) and we describe it briefly here. The spatial propagation of relativistic electrons injected by AGNs is followed in postprocessing using a Lagrangian advection algorithm, as detailed in Wittor et al. (2020). The spectral energy evolution is modeled by numerically integrating Fokker–Planck equations of an initial power-law distribution of electron momenta, N(γ) ∝ γ⁻², in the range of ${\gamma }_{\min }=50$ and ${\gamma }_{\max }={10}^{5}$ (where γ is the Lorentz factor of electrons). We consider continuous energy transfer via synchrotron emission, inverse Compton scattering, Coulomb collisions, adiabatic compression/expansion, and injection and reacceleration of electrons from shock induced DSA (e.g., Kang et al. 2012). The AGNs produced by such a model are compatible with FR II morphologies and start with a radio luminosity of ∼10⁴⁰–10⁴¹ erg s⁻¹ Hz⁻¹ at 120 MHz. The lobes of these AGNs spatially evolve by diffusing into the ICM and rapidly fade in the radio band due to adiabatic losses and radiative cooling, except when perturbations in the ICM reaccelerate fossil electron spectra via adiabatic compression and, most importantly, DSA processes. Turbulent reacceleration processes have not been included.

By visually inspecting the radio emission maps of all snapshots in the simulation, we identified a short (∼100–250 Myr) evolutionary stage, some 2.2 Gyr after the original AGN outburst, during which the aging lobes released by three of the four AGNs produced a radio morphology that loosely resembles the USS Jellyfish in extent, total power, and radio spectral slope. ¹⁴ In the epochs prior, the AGN lobes rapidly dimmed; however, a confluence of factors causes a transient rebrightening coinciding with the mixing of three of the four ancient lobes at the same time as the passing of several weak shocks (${ \mathcal M }\lt 2$) triggered by both cluster growth and interactions of the AGN lobes themselves. The top panel of Figure 7 shows the contours of radio emission at 120 MHz from this snapshot if the source were placed at the same distance as A2877, with contours marking ≥7 mJy beam⁻¹ regions, while the colors give the spectral index in the 87–150 MHz frequency range. At least qualitatively, the detectable emission in the simulated emission map resembles the complex shape of the real USS Jellyfish, including its elongated threads and radio substructures in the “head.” The spectral index map has a rather uniform distribution with α ∼ −3, with some patches of steeper emission in the threads. This model takes into account the role of weak shock (re)acceleration in the ICM, driven by matter accretion events that took place in the host cluster after the injection of jets. The middle panel shows the same snapshot but without “cluster weather”—that is, excluding all ICM interactions, such as weak shocks induced by gas motions associated with cluster growth—and points to the necessity in this particular case for these external mechanisms to draw out the tentacles and filamentary structures. The bottom panel shows the dark matter density, which, as a proxy for the location of galaxies in the simulation, shows that three of the peaks in the radio emission still closely align with their host galaxies, even after significant evolution of the system.

Zoom In Zoom Out Reset image size

This simulation tentatively suggests that multiple aging and interacting AGN plasmas, alongside weak shock-induced adiabatic compression and DSA, are a plausible and sufficient explanation for diffuse and very steep cluster emission like the USS Jellyfish, and that alternative mechanisms, such as those developed to explain the GReET, are not necessarily required here. The simulation also shows the potential rarity of the scenario due to both its short-lived, transient nature in a lengthy 2.2 Gyr evolution and the requirement for an otherwise radio-quiet cluster across the remainder of that evolution.

We have reported on the discovery of the steepest-spectrum synchrotron source to date, the USS Jellyfish, which lies to the northwest of the cluster A2877. We have argued that the source is a polyphoenix: that is composed of multiple aged and mixed AGN populations reaccelerated by a common large-scale event. The currently available X-ray data do not show the presence of cold fronts, sloshing, or other shock systems in the cool-core cluster; however, the quality of these data likely means that weak shocks are undetectable. We have also presented recent magnetohydrodynamic simulations showing that a combination of weak shocks inducing standard DSA and adiabatic compression and acting upon ancient interacting AGN plasmas are capable of producing diffuse, USS emission akin to the USS Jellyfish without recourse to other, more exotic (re)acceleration mechanisms. These same simulations show that the USS Jellyfish may also be a short-lived, transient phase in the evolution of the system.

Follow-up work on A2877 should make high-sensitivity and high-resolution X-ray observations of A2877 a priority in an effort to discern the presence and nature of any shocks in the northwest and detect telltale signs of cool-core sloshing that may be present. Additionally, both higher-resolution and higher-sensitivity radio observations of the USS Jellyfish may provide us with a better picture of its morphology; however, the incredible steepness of its spectrum makes follow-up observations above 300 MHz unlikely to detect any of the extended emission. We may have to wait for the development of SKA Low for a higher-resolution, low-frequency telescope able to observe so far south.

The authors thank Benjamin Quici for helpful discussions during the preparation of this manuscript. This scientific work makes use of the Murchison Radio-astronomy Observatory, operated by CSIRO. We acknowledge the Wajarri Yamatji people as the traditional owners of the observatory site. Support for the operation of the MWA is provided by the Australian Government (NCRIS) under a contract to Curtin University administered by Astronomy Australia Limited. We acknowledge the Pawsey Supercomputing Centre, which is supported by the Western Australian and Australian Governments. The Australia Telescope Compact Array is part of the Australia Telescope National Facility, which is funded by the Australian Government for operation as a National Facility managed by CSIRO. F.V. acknowledges financial support from the ERC Starting grant MAGCOW, No. 714196. The cosmological simulations in Section 4 were performed with the ENZO code (http://enzo-project.org) under projects “stressicm” and “hhh44” at the Jülich Supercomputing Centre (PI: F. Vazza). D.W. is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) 441694982.