Powerful AGN jets and unbalanced cooling in the hot atmosphere of IC 4296

Abstract

We present new Karl G. Jansky Very Large Array (VLA, 1.5 GHz) radio data for the giant elliptical galaxy IC 4296, supported by archival radio, X-ray (Chandra, and XMM–Newton) and optical (SOAR, and HST) observations. The galaxy hosts powerful radio jets piercing through the inner hot X-ray emitting atmosphere, depositing most of the energy into the ambient intracluster medium (ICM). Whereas the radio surface brightness of the A configuration image is consistent with a Fanaroff–Riley Class I system, the D configuration image shows two bright, relative to the central region, large (⁠|${\sim } 160\, \rm {kpc}$| diameter), well-defined lobes, previously reported by Killeen et al., at a projected distance |$r\gtrsim 230\, \rm {kpc}$|⁠. The XMM–Newton image reveals an X-ray cavity associated with one of the radio lobes. The total enthalpy of the radio lobes is |${\sim }7\times 10^{59}\, \rm {erg}$| and the mechanical power output of the jets is |${\sim } 10^{44}\, \rm {erg\, s}^{-1}$|⁠. The jets are mildly curved and possibly rebrightened by the relative motion of the galaxy and the ICM. The lobes display sharp edges, suggesting the presence of bow shocks, which would indicate that they are expanding supersonically. The central entropy and cooling time of the X-ray gas are unusually low and the nucleus hosts a warm Hα + [|$\rm{N\,{\small II}}$|] nebula and a cold molecular CO disc. Because most of the energy of the jets is deposited far from the nucleus, the atmosphere of the galaxy continues to cool, apparently feeding the central supermassive black hole and powering the jet activity.

galaxies: active, galaxies: ISM, X-rays: galaxies

1 INTRODUCTION

The oldest, largest known galactic structures in the Universe are giant elliptical galaxies. These systems are evolving contrary to the standard hierarchical galaxy formation scenario (Thomas, Maraston & Bender 2002). Observations and theoretical models indicate that they formed around 0.2–1 billion years after the big bang, went through a fast star-forming phase, and afterward grew only by dry mergers (Thomas et al. 2010; Onodera et al. 2015).

Initially, the evolution of giant ellipticals is dominated by dark matter, which clumps into haloes via gravity. When the mass of a halo reaches ∼10¹² M_⊙ (Correa et al. 2018), the inflowing gas passes through accretion shocks, and its temperature increases to several million Kelvin, forming an X-ray emitting atmosphere (further replenished by stellar mass loss; Pellegrini et al. 2018) and quenching star formation (Cattaneo et al. 2009). These galaxies will continue to grow passively by interactions and dry mergers with other galaxies (Cattaneo et al. 2009). On the other hand, the sub-Eddington accretion rate of the gas onto the central supermassive black hole enables the so-called radio-mechanical (or radio mode) active galactic nucleus (AGN) feedback, thereby producing jets and lobes of relativistic plasma, visible in the radio band, which are able to propagate well outside their galaxy hosts.

Radio wavelength studies of giant elliptical galaxies are important for answering outstanding questions about the heating and cooling of their X-ray emitting atmospheres and about the physics of their AGNs (Brighenti & Mathews 2006). In clusters of galaxies, there is strong observational evidence for radio-mode AGN feedback balancing the cooling of the intracluster medium (ICM, e.g Churazov et al. 2000; Fabian et al. 2003; Bîrzan et al. 2004; Rafferty et al. 2006; Brüggen & Kaiser 2002) and preventing cooling flows and dramatic star formation (see Fabian 1994). It is believed that radio-mechanical feedback plays a critically important role also at smaller scales, in preventing atmospheric cooling in massive galaxies (Werner et al. 2019).

In the case of giant ellipticals, the so-called cooling flow problem is more severe. Shorter cooling time and larger amount of gas returned by the stellar populations over the lifetime of the galaxy (Cattaneo et al. 2009) place stronger demands on the AGN feedback as a source of heat balancing the radiative cooling. In some cases, the high-velocity radio jets are able to drill through the hot galactic atmospheres without significantly affecting the host galaxy, which may lead to unbalanced cooling in the innermost part of the galaxy (Sun et al. 2005b). Comprehensive studies of such systems are essential for our understanding of the role of AGN feedback in galaxies, galaxy groups, and galaxy clusters.

The powerful AGN jets penetrating the innermost hot galactic atmosphere and depositing most of their energy out at radii |$r\gtrsim 230\, \rm {kpc}$| make the giant elliptical galaxy IC 4296, in the central region of the galaxy cluster Abell 3565, a source of high interest. Its radio counterpart, PKS 1333−33, has attracted the attention of astronomers for over 30 yr. Historically, the first comprehensive studies were published by Killeen, Bicknell & Carter (1986a), Killeen, Bicknell & Ekers (1986b), and Killeen & Bicknell (1988) using data from Very Large Array (VLA) at 1.3, 2, 6, and 20 cm. In particular, Killeen et al. (1986a) analysed X-ray data of the system from the imaging proportional counter (IPC) onboard the Einstein observatory. In addition to prominent jets extending over 35 arcmin, the authors reported a barely resolved diffuse X-ray source coinciding with the position of the galaxy.

In this paper, we present new radio data of IC 4296 obtained with the VLA in the A and D configurations at 1.5 GHz and compare them with archival observations available at lower frequencies in the TIFR GMRT Sky Survey (TGSS; Intema et al. 2017) and in the GaLactic and Extragalactic All-sky Murchison Widefield Array(GLEAM; Hurley- Walker et al. 2016).

While our paper presents new, deep radio observations, our goal is to study this system in a more comprehensive way. Therefore, we also analyse archival X-ray data from the Chandra and XMM–Newton observatories, as well as narrow-band Hα + [|$\rm{N\,{\small II}}$|] images from the 4.1 m Southern Astrophysical Research (SOAR) telescope and optical data from the Hubble Space Telescope (HST). This allows us to study the interaction between the jet/lobes and the surrounding hot and cold thermal gas phases.

Our paper is organized as follows: Section 2 includes the observational details and data reduction methods. Our results are presented and described in Section 3. A thorough discussion in Section 4 addresses the most important outcomes of our multiwavelength analysis connecting all these multiwavelength observations, with a particular emphasis on the relation between the radio and X-ray band. Finally, our conclusions and remarks are summarized in Section 5.

Throughout this paper, we used the following cosmological parameters: |$\rm H_{0}=69.6\, \rm {km\, s}^{-1}\, \rm {Mpc}^{-1}$| (Bennett et al. 2014), Ω_M = 0.286, and Ω_Λ = 0.714. At the redshift of 0.0125, the distance of IC 4296 is |$49\, \rm {Mpc}$| (Mei, Silva & Quinn 2000), estimated using the surface brightness fluctuation method, with the corresponding scale of |$0.256\, \rm {kpc}\, \rm {arcsec}^{-1}$|⁠. The spectral index α of the synchrotron radiation as a function of the flux density S_ν at the frequency ν is defined as S_ν ∝ ν^α. The elemental abundances are expressed with respect to the proto-solar values from Lodders, Palme & Gail (2009).

2 OBSERVATIONS AND DATA ANALYSIS

2.1 Radio observations and analysis

IC 4296 was observed on the 2015 July 11 by the Karl G. Jansky VLA in A configuration (project code: 15A-305; PI: Werner) and on the 2018 October 26 (project code: 18A-317, PI: Grossova) in D configuration. The L-band receiver covering the frequency range spanning from 1 to 2 GHz is divided into 16 spectral windows with 64 1000 kHz wide channels. The one-hour long observations consist of ≈40 min integration time on the target and ≈10 min on both the standard VLA flux density calibrator 3C 286 and a nearby compact source J1316−3338 used as amplitude and phase calibrator. The VLA data were reduced and imaged using the National Radio Astronomy Observatory (nrao) pipeline Common Astronomy Software Applications (casa, McMullin et al. 2007), v4.7.2 and v5.0.0 for A and D configuration data, respectively.

A careful approach was followed to flag radio frequency interferences (RFI). First, we used the auto-flagging algorithm tfcrop in casa to identify the outliers in the time–frequency plane. Then, we used the Offringa’s RFI software aoflagger (Offringa, van de Gronde & Roerdink 2012). The combination of the two strategies, for the A configuration, flagged in total 46 per cent of the data, with spectral windows 2, 3, 8, and 9 almost entirely flagged. In case of the D configuration, more than half of the data had to be flagged.

Afterwards, we used a model for the flux calibrator 3C 286 provided by the casa package to set up the flux density scale determined by Perley & Butler (2013). The initial gain calibration was performed for central channels on both calibrators to average over the small variations of phase with time in the bandpass. Next, we performed the bandpass calibration. The relative delays of each antenna in comparison with the reference antenna were derived and complex bandpass solutions were calculated to include variations in gain with frequency. The next step determined complex gains for both calibrators. Finally, the derived calibration solutions were applied to the target.

The deconvolution and imaging were performed by the casaclean algorithm in the multifrequency synthesis mode. We produced images over a wide range of resolutions and with different uv-tapers and weighting schemes to emphasize the radio emission on various spatial scales. A second-order Taylor polynomial (nterms = 2, McMullin et al. 2007) was used, to take into account the spectral behaviour of bright sources. To refine the calibration for the target, we performed two cycles of phase and one cycle of amplitude and phase self-calibration.

We also constructed a spectral index map using the VLA D configuration image at 1.5 GHz and a GLEAM radio image at 158 MHz. To create a reliable spectral index map it is essential to compare the images using the same resolution and the same pixel size. First, we deconvolved the 2D VLA calibrated data using the uvtaper option within the casaclean algorithm to create a total flux intensity image close to the resolution of the GLEAM image. Then, we smoothed the image with the exact GLEAM restoring beam size¹ using the casa task imsmooth. Secondly, the GLEAM image was regridded to get the same pixel size as the VLA D configuration image. We used the casaimmath task, to calculate flux densities, spectral indices, and their errors following Rajpurohit et al. (2018). We assumed flux density uncertainties (f_err) of |$4{{\ \rm per\ cent}}$| for the VLA data (Perley & Butler 2013) and |$8{{\ \rm per\ cent}}$| for the GLEAM data (Hurley-Walker et al. 2017).

2.2 Archival X-ray observations

2.2.1 Archival Chandra observations and analysis

IC 4296 was observed by the Chandra Advanced CCD Imaging Spectrometer in the S-array (ACIS-S) on the 2001 September 10 (Obs. ID: 2021) and the 2001 December 15 (Obs. ID: 3394) with a total clean exposure time of 48 ks. The on-axis point spread function (PSF) is better than 1arcsec for Chandra. The data reduction and analysis are described in detail in Lakhchaura et al. (2018).

2.2.2 Archival XMM–Newton observations and analysis

In addition to Chandra, we also used an archival XMM–Newton observation (Obs. ID: 0672870101) of IC 4296, which was performed on the 2011 July 11. We took advantage of both the EPIC (MOS 1, MOS 2, and pn) and RGS (RGS 1 and RGS 2) instruments to fully investigate this system. The on-axis PSF of the X-ray telescopes on XMM–Newton is of the order of 10 arcsec.

The entire EPIC (European Photon Imaging Camera) data reduction procedure and EPIC imaging extraction were done using the XMM Science Analysis System (sas) software v17.0.0 following Mernier et al. (2015) to which we refer for further details. After filtering the EPIC data from flaring events, the cleaned exposures are 46, 47, and 44 ks for MOS 1, MOS 2, and pn, respectively.

The EPIC spectral analysis reported in Section 3.2 follows the general prescription detailed in Mernier et al. (2015). We used the spectral fitting package spex v3.04 (Kaastra, Mewe & Nieuwenhuijzen 1996; Kaastra et al. 2017) to analyse the obtained EPIC spectra. Since the signal-to-noise ratio in the extracted EPIC regions is much weaker than in the central regions analysed with ChandraACIS-S (Section 3.2), a careful modelling of the background is necessary here. This modelling, taking into account the galactic thermal emission, the local hot bubble, the unresolved point sources (or cosmic X-ray background), the quiescent particle background, and the residual soft-proton background is also fully described in Mernier et al. (2015), and references therein. Finally, we modelled the X-ray emission of IC 4296 with a redshifted and absorbed collisional ionization equilibrium plasma model (cie), in which the abundances are fixed to 0.3 proto-solar (Urban et al. 2017). The MOS 1, MOS 2, and pn spectra were fitted simultaneously and the spectra were deprojected using the dsdeproj tool (Russell, Sanders & Fabian 2008). The free parameters are thus the temperature (kT) and the spex normalization.

The RGS data reduction was performed using the sas task rgsproc, which also takes care of filtering the data from flared events and of estimating the background from the RGS CCD 9 (where no source counts are expected). The first-order RGS 1 and RGS 2 raw spectra, extracted within 0.8 arcmin (i.e. 90 per cent of the PSF) in the cross-dispersion direction, are then background-subtracted and combined using the sas task rgscombine. Compared to EPIC, the RGS spectra are in principle much better suited for investigating the heating–cooling balance in the central X-ray atmosphere of IC 4296 (Section 3.2). Therefore, in addition to the cie emission, we model our combined RGS spectrum with a classical ‘cooling-flow’ (cf) component. The multiplicative component lpro (calculated from the MOS 1 detector image in the 0.3–2 keV band) is also applied to account for the instrumental broadening of the lines due to the spatial extent of the source. The lower and higher temperature limits of the cf model are respectively fixed to 0.1 keV and tied to the temperature of the cie model. The free parameters are the normalization of the cie component, the cooling rate of the cf component, and the O, Ne, and Fe abundances. The other abundances are tied to that of Fe, with the exception of Mg which is tied to the O abundance.

2.3 Archival optical observations and analysis

Narrow-band Hα +[|$\rm{N\,{\small II}}$|] imaging of IC 4296 was obtained by Sun et al. (in prep.) with the 4.1 m SOAR telescope using the SOAR Optical Imager (SOI). The full description of these observations is given in the survey paper (Sun et al. in preparation), but we briefly summarize them here. IC 4296 was observed on the 2015 July 21 for 1800 s in both on (6649/76) and off (6737/76) band filters.² Basic reduction was performed using SOI-specific iraf scripts based on the mscred package.

In addition to the SOAR image, we used archival HST Imaging with the Advanced Camera for Surveys (ACS) High Resolution Channel (HRC) obtained as part of program GO-9838 (PI: L. Ferrarese) observed on the 2004 August 07. ACS/HRC images are in the F435W filter and the FR656N filter, which was tuned to a central wavelength of |$6644.3\, \rm {\mathring{\rm A}}$|⁠; this tuning samples Hα at the redshift of IC 4296. These observations were previously discussed by Dalla Bontà et al. (2009) and presented also in Lauer et al. (2005) and Balmaverde & Capetti (2006). We did not continuum-subtract the ACS/HRC images, but a strong nuclear source is seen in the FR656N image that is not seen in the corresponding F435W image. As was discussed by Dalla Bontà et al. (2009), there is strong Hα +[|$\rm{N\,{\small II}}$|] emission that extends beyond the central dust disc visible in the right-hand panel of Fig. 8 (for more details see also Section 3.3).

3 RESULTS

3.1 Radio morphology

The 1.5 GHz VLA A configuration radio image is shown in Fig. 1 (left-hand panel) with a resolution of |$3.94\, \rm {arcsec}\times 3.26\, \rm {arcsec}$| and a root-mean-square (rms) noise of 0.06 mJy beam⁻¹. The total intensity image clearly reveals a bright nucleus and extended emission in two jets, with several knots on both sides, terminating in diffuse radio plumes. The two well collimated and almost symmetric jets emit via synchrotron radiation in the central regions. However, interestingly, the jets show noticeable rebrightening further away from the nucleus at |$r\sim 10\, \rm {kpc}$|⁠, where the radio plumes broaden, as mentioned previously by Killeen & Bicknell (1988). Based on our VLA data obtained in A configuration (which is less sensitive to extended structures), the total extent of the radio structure is |${\sim }120\, \rm {kpc}$| with the northwestern jet extending out to |$r\sim 70\, \rm {kpc}$| and the southeastern to |$r\sim 50\, \rm {kpc}$|⁠. The total integrated flux density of the source in A configuration is |$3.98\pm 0.40\, \rm {Jy}$|⁠. IC 4296 shows a gradual decrease in surface brightness from the centre, which is typical of Fanaroff–Riley (FRI) radio galaxies.

Figure 1.

The VLA total intensity images of IC 4296 at 1.5 GHz. Left-hand panel: the A configuration image shows that the surface brightness of jets decreases as a function of radius, although we also see rebrightening in the radio plumes at |${\sim }10\, \rm {kpc}$| from the nucleus. The restoring beam size is |$3.94\, \rm {arcsec}\times 3.26\, \rm {arcsec}$| and is shown in the bottom right corner. Right-hand panel: the D configuration image revealed well-confined radio lobes extending up to almost 300 kpc from the central regions of IC 4296, previously reported by Killeen et al. (1986b). The restoring beam is |$90.15\, \rm {arcsec}\times 23.02\, \rm {arcsec}$| and is shown in the bottom right corner. The contour levels are at [−1,1,2,4,8,16] × 5 rms noise, with corresponding rms noise of |$0.06$| and of |$0.2\, \rm {mJy\, beam}^{-1}$| for VLA A and D configuration data, respectively.

The low-resolution D configuration image is shown in Fig. 1 (right-hand panel). The image has a resolution of |$90.15\, \rm {arcsec}\times 23.02\, \rm {arcsec}$| and the rms noise level is |${\sim } 0.2\, \rm {mJy\,beam^{-1}}$|⁠. The total integrated flux density of IC 4296 in D configuration is |$8.30\pm 0.83\, \rm {Jy}$|⁠. The difference between the extracted integrated flux density values in A and D configurations lies in the fact that the A configuration data with the missing short baselines are not able to recover the highly peaked signal from the more extended diffuse emission in the jets and lobes. Two well-confined lobes are visible in the D configuration image at a projected distance of |${\sim }230$| and 290 kpc from the bright nuclear region, in the northwestern and southeastern directions, respectively.

Since the plumes and/or lobes in FR I and FR II radio galaxies tend to have steep radio spectra, we also inspected the archival TGSS radio maps at 153 MHz. These revealed a radio structure close to that observed at 1.5 GHz in our VLA A configuration image, but also a weak diffuse lobe-like emission extending far beyond the A configuration image (black contours in Fig. 2). The GLEAM data at 154–162 MHz also confirm the two extended lobe-like structures (white contours in Fig. 2).

Figure 2.

The VLA D configuration radio map at 1.5 GHz overplotted with the TGSS radio contours at 153 MHz (in black), VLA A configuration contours at 1.5 GHz (in red), and the GLEAM radio maps at 154–162 GHz (in white). Contours levels were created at |${}[-1, 1, 2, 4, 8, 16]\times 5\,$| rms noise for the TGSS, GLEAM, VLA A configuration data with corresponding rms noise of |$3$|⁠, |$0.12$|⁠, and |$0.06\, \rm {mJy\, beam}^{-1}$|⁠.

The spectral index map created using the GLEAM data at 158 MHz and the VLA D configuration data at 1.5 GHz is shown in Fig. 3. As expected, the central regions of the galaxy show a flat spectral index of −0.5 and a steepening is present in the lobes. They both have steep spectra with indices of −0.7 and −1.1, respectively (Fig. 3). The spectral index map appears uniform over the two lobes in agreement with the lack of clear hotspots that would have flatter radio spectra. The uncertainties on the spectral index are of the order of a few per cent.

Figure 3.

The spectral index map for IC 4296 reveals a flat spectral index in the central region of the jet, whereas in the two lobes we observe a steep spectral index of −0.7 and −1.1. The map was created using the 158 GHz GLEAM and the 1.5 GHz VLA D configuration image at |$155\, {\rm {arcsec}}\times 143\, \rm {{arcsec}}$| resolution (the beam is shown in the right-hand corner of the map) and it is overlaid by VLA contours corresponding to the VLA smoothed image, in respect to GLEAM image to reach the same beam size. The rms noise reached value of |${\sim }1.5\, \rm {mJy\, beam}^{-1}$|⁠. Contour levels are |${}[-1, 1, 2, 4, 8, 16]\times 5\, {\mathrm{rms\, noise}}$|⁠.

The radio power of |${\sim }10^{24}\rm {\, W\, Hz}^{-1}$| at 1.4 GHz³ obtained from casa total intensity D configuration image is close to the boundary radio power of the FR dichotomy of |${\sim }10^{24.5}\rm {\, W\, Hz}^{-1}$| at 1.4 GHz (Owen & Ledlow 1994). With its absolute R-band magnitude of |$M_{\mathrm{R}}=-23.33\pm 0.24\, \rm {mag}$|⁠, IC 4296 is placed below the boundary division line of the Ledlow–Owen diagram, where most of the FRI sources are located.

3.2 Comparison with the X-ray data

The characteristic radius of the X-ray atmosphere in Fig. 4 (left-hand panel) observed by ChandraACIS-S (see also Pellegrini et al. 2003) is much smaller than the total extent of the radio jets. The jets emanating from the nuclear region are at first well collimated and, after piercing through the innermost parts of the hot X-ray emitting atmosphere, at |$r\sim \, 10\, \rm {kpc}$|⁠, they widen significantly, as previously noticed by Killeen & Bicknell (1988).

Figure 4.

Left-hand panel: the Chandra X-ray image at 0.5–5 keV overlaid with white contours of our VLA image at 1.5 GHz obtained in A configuration with same contour levels as defined in Fig. 1. The image reveals an excess of the X-ray emission, marked by a red circle, to the southwest of IC 4296. Right-hand panel: the background-subtracted, exposure-corrected XMM–Newton EPIC image (combining MOS 1, MOS 2, and pn) in the 0.3–2 keV band overlaid with contours of the VLA D configuration image same as in Fig. 1 (right-hand panel). The X-ray image reveals a decrement, a likely cavity, corresponding to the northwestern radio lobe. The overplotted red partial annuli indicate the spectral extraction regions used to determine the temperature and the density of the ICM outside of the northern radio lobe.

Lakhchaura et al. (2018) investigated the radial profiles of the entropy and the cooling time (t_cool) over free-fall time (t_ff) ratio. The cooling time and entropy decrease with decreasing radius down to |$\lt 1\, \rm {kpc}$|⁠. The power-law fit of the radial entropy profile reveals a steep slope with an index of ∼1.2. Even below 2 kpc, the slope of the entropy profile remains steep with an index of ∼1.0, and in the innermost radial bin the entropy reaches a value of |$K_{0}\sim 1.5\, \rm {keV\, cm^2}$| (see the appendix in Lakhchaura et al. 2018). In fact, IC 4296 has the steepest entropy profile of all galaxies in the sample of Lakhchaura et al. (2018), followed by that of NGC 4261 (see Fig. 5). This will be discussed further in Section 4.2.

Figure 5.

The entropy profiles of IC 4296 and NGC 4261 in comparison with the sample of Lakhchaura et al. (2018). The red, green, and blue solid lines show median profiles for cool gas free, nuclear cool gas, and extended cool gas systems, respectively and the shaded regions show the median absolute deviation spreads about the medians. The entropy profiles of IC 4296 and NGC 4261 show much steeper trends than the other galaxies in the sample.

The cooling time in the centre of IC 4296 is |$t_{\rm cool}\lt 11.2\, \rm {Myr}$|⁠. The ratio of the cooling time to free-fall, time t_cool/t_ff, reaches a value close to 10 in the centre of the galaxy, where the atmosphere most likely continues to cool, feeding the central supermassive black hole and powering the jets.

The archival Chandra X-ray image reveals an excess of X-ray emission in the southwestern part of the structure, in between the northwestern and southeastern sides of the jet, as marked by a red circle in the Chandra X-ray image in Fig. 4 (left-hand panel). The detection significance of such X-ray diffuse emission is above |$5\, \sigma$|⁠. Surface brightness profiles derived along this extended X-ray emission do not reveal any sharp edges.

Our XMM–Newton EPIC image extracted in the 0.3–2.0 keV band reveals a cavity with a diameter of |${\sim }160\, \rm {kpc}$| associated with the northwestern radio lobe (see Fig. 4; right-hand panel). We find a small misalignment between the radio lobe and the X-ray cavity. It could be the result of projection effects, where the part of the cavity closer to the mid-plane of the source is clearly visible in the X-ray image, while the part associated with the visible radio emission might be offset from the mid-plane and thus not seen clearly in the images.

We extracted and fitted the deprojected spectra from four partial spherical shells in the direction of the northwestern jet (traced by the red sectors in Fig. 4; right panel). The corresponding radial profile of the electron density (n_e) is shown in Fig. 6. We have subtracted the emission contribution of the gas at larger radii (i.e. beyond our field of view), which we estimated by extrapolating a beta model fit to our initial density profile. The temperature and total density (electron plus ions) of the gas surrounding the cavity (i.e. the last radial bin in Fig. 6) are respectively kT =0.95 ± 0.01 keV and n =9.6 ± 0.05 × 10^-4cm^-3, which correspods to a pressure of nkT =9.12 × 10^-4 keV cm^-3. Assuming the northwestern cavity is spherical with a radius of 80 kpc, the 4pVwork performed by the jet to displace the ICM is ∼ 3.7 × 10⁵⁹erg. Assuming a similar enthalpy for the southeastern cavity, which is outside of the XMM-Newtonfield-of-view, we obtain a total enthalpy of ∼ 7 × 10⁵⁹erg.

Figure 6.

Deprojected electron density profile for the gas surrounding the northwestern cavity of IC 4296. The two spectral extraction regions, whose values have been combined and averaged in this figure, are shown by the red sectors in Fig. 4 (right-hand panel).

The combined XMM–Newton RGS 1 + RGS 2 spectrum is shown in Fig. 7. Some lines (e.g. Fe xvii at 15 and 17 Å, O viii at 19 Å, rest frame) are clearly identified as they are typical of a cool (≲1 keV) gas. Our best-fitting model provides a temperature of |$0.79 \pm 0.05\, \rm {keV}$| and an Fe abundance of 0.13 ± 0.02 for the cie component, as well as a cooling rate of 4.5 ± 1.0 M_⊙ yr⁻¹ for the cf component. In this spectrum of limited quality, however, the cooling rate and Fe abundance parameters seem to share some degeneracy, with our inferred cooling rate dropping to 2.1 ± 0.6 M_⊙ yr⁻¹ when fixing the abundances to 0.3 proto-solar. This is further discussed in Section 4.2.

Figure 7.

Combined first-order RGS 1 + RGS 2 spectrum of IC 4296, extracted within 0.8 arcmin in the cross-dispersion direction. Our best-fit model (cie + cf, see the text) is overplotted in red.

3.3 Comparison with the optical data

The narrow-band SOAR Hα + [|$\rm{N\,{\small II}}$|] image shown in Fig. 8 (right-hand panel) indicates that the ionized gas is circumnuclear and the ionized disc is perpendicular to the jets. The SOAR data also provide a hint for entrainment along the northwestern radio plume. The HST data show a warped, dust disc with a |$0.9\, \rm {arcsec}$| radius (see Fig. 8; left-hand panel), which does not appear to be perpendicular to the jets (see Schmitt et al. 2002, for more details) and is visible on smaller scales than the emission in the SOAR image (Fig. 8; right-hand panel). Interestingly, recent studies of CO emission for a sample of giant elliptical galaxies by Boizelle et al. (2017) and Ruffa et al. (2019) revealed the presence of a circumnuclear molecular gas disc, which is cospatial with the dust (Ruffa et al. 2019).

Figure 8.

Left-hand panel: the HST image of a dust disc in IC 4296. The narrow-band Hα image (in green) was obtained using ramp filter adjusted to the redshift of the galaxy. Two white arrows represent jets emitted from the central nuclear regions. Right-hand panel: the narrow-band Hα + [|$\rm{N\,{\small II}}$|] image from the SOAR telescope overlaid with the VLA A configuration radio contours.

4 DISCUSSION

4.1 The nature of the radio source

Our new VLA data in D configuration at 1.5 GHz confirm the presence of the two lobes previously detected by Killeen et al. (1986b). While the extended emission resembles the radio morphology of FR II radio galaxies, the lobes lack compact hotspots. The total extent of the radio structure is |${\sim }500\, \rm {kpc}$|⁠, which is in agreement with Killeen et al. (1986b). The compactness of the radio lobes relative to their distance from the galaxy core, make IC4296 very similar to other powerful radio galaxies.

From the X-ray analysis, we conclude that it would take |$1.5\times 10^{8}\, \rm {yr}$| to inflate the northwestern X-ray cavity with a radius of |${\sim }80\, \rm {kpc}$| at the sound speed of |$c_{\rm s}=516\rm {\, km\, s}^{-1}$| for |$kT=1\, \rm {keV}$| gas. Assuming this expansion time, we obtain a jet power of 1.6 × 10⁴⁴ erg s⁻¹. Another estimate considers the sound-crossing time of the projected distance of the cavity from the centre of the galaxy of |$2.2\times 10^{8}\, \rm {yr}$| resulting in a jet power of 10⁴⁴ erg s⁻¹.

The radio morphology of the source indicates possible supersonic expansion. In the boundary layer, where the well-defined northwestern radio lobe meets the ambient ICM (Fig. 1; right-hand panel), the surface brightness is increased and the end of the jet forms a lobe with a sharp-edged morphology. According to 3D hydrodynamic simulations by Massaglia et al. (2016) and Perucho & Martí (2007), this sharp head of the jet is consistent with bow shocks, which are typical indicators of the supersonic expansion relative to the surrounding medium. The inconsistency with the classical FR I class, low power radio sources with subsonic expansions, lies in the rebrightenings and sharp-edged structure of the northwestern radio lobe suggesting its supersonic expansion.

The radio morphology of IC 4296 shows similarities with the morphology of 3C 348 (Hercules A, often classified as an FR I/II radio source, Siebert et al. 1996; Gizani & Leahy 2003), which is one of the most luminous radio sources with a total power comparable to the Cygnus A galaxy (Gizani, Cohen & Kassim 2005). Hercules A also has missing compact hotspots and possesses an unusual morphology dominated by jets and inflated steep spectrum lobes.

The jets, piercing through the galactic atmosphere and depositing their energy into the surrounding ICM, are also present in the FRI radio source NGC 4261 (a.k.a. 3C 270, O’Sullivan et al. 2005). Interestingly, we find that the derived enthalpy for the cavities/lobes of IC 4296 is ∼30 times larger than for NGC 4261 (O’Sullivan et al. 2011), mainly due to the size of the radio lobes (∼6 times larger in the case of IC 4296). As further discussed in Section 4.2, in these two cases the central regions also show steep entropy (see Fig. 5) and cooling time profiles (Werner, Allen & Simionescu 2012; Voit et al. 2015b) and a prominent nuclear dust disc fuelling the AGN closely related to powerful radio jets (Jaffe et al. 1993; Jaffe & McNamara 1994), which might have cooled from the hot X-ray emitting atmosphere of the galaxy. Such a dust and molecular disc are also present in the brightest cluster galaxy of Hydra A, which also displays powerful jet activity (McNamara 1995; Taylor 1996).

4.2 The unbalanced cooling in IC 4296

The X-ray image overlaid with the radio contours, in Fig. 4, indicates that while in the inner part of the galaxy, at |$r\lesssim 10\, \rm {kpc}$|⁠, the radio jets are well collimated by the thermal pressure of the hot atmosphere, their width increases at larger radii. Nevertheless, these powerful radio jets pierce through the inner hot X-ray emitting atmosphere of the galaxy and extend to radii over r ≳ 230 kpc, where they deposit their energy into the outer galactic atmosphere and the ambient ICM.

Similarly to NGC 4261, the entropy profile of IC 4296 (Fig. 5) shows a steeper slope than for all the other giant elliptical galaxies analysed by Lakhchaura et al. (2018). In fact, it appears to be decreasing all the way to |$r\lt 1\, \rm {kpc}$| (see also Panagoulia, Fabian & Sanders 2014; Babyk et al. 2018). At |$r\lt 1\, \rm {kpc}$|⁠, t_cool/t_ff drops to ≈10 and the gas may become thermally unstable (Sharma et al. 2012; Gaspari, Ruszkowski & Sharma 2012; Kunz et al. 2012; McCourt et al. 2012; Gaspari, Ruszkowski & Oh 2013; Voit et al. 2015a). These results suggest that, whereas the expanding radio jets and lobes may be able to heat the surrounding gas at r ≳ 10 kpc, the inner, well collimated, part of the jets do not appear to be depositing sufficient heat to the innermost rapidly cooling X-ray atmosphere.

At first glance, this picture is in line with our analysis of the XMM–Newton RGS spectra, as they are consistent with an appreciable mass deposition rate of |$4.5\pm 1.0\, \mathrm{M}_\odot \, \rm {yr}^{-1}$|⁠. This rate is particularly high compared to other elliptical galaxies and groups whose RGS spectra reveal mass deposition rates typically less than 1 M_⊙ yr⁻¹ (Liu et al.2019). We caution, however, that this measurement suffers from large systematic uncertainties because the limited quality of the data implies degeneracies between some parameters. For example, fixing all the abundances to 0.3 proto-solar (which is more realistic than our best-fitting 0.13 proto-solar value; e.g. Urban et al. 2017) yields a lower mass deposition rate of |$2.1\pm 0.6\, \mathrm{M}_\odot \, \rm {yr}^{-1}$|⁠. Moreover, we cannot exclude even higher abundances (i.e. close to the proto-solar value) in the inner ∼10 kpc region, which would then further revise lower the inferred cooling rate. Future deeper observations will thus be important to confirm (or disprove) the unusually high cooling rate in this source. Assuming nevertheless our initial deposition rate and that the bulk of this material accretes onto the central supermassive black hole and is converted into jets with a 10 per cent efficiency, it will result in a jet power of |${\sim } 2.6\times \, 10^{46}\, \rm {erg\, s^{-1}}$|⁠. This is more than two orders of magnitude higher than the actual jet power estimated in Section 4.1 (⁠|${\sim } 10^{44}\, \rm {erg\, s^{-1}}$|⁠). We therefore conclude that the currently assumed rate of cooling is largely sufficient to effectively produce the observed radio jets.

In the centre of the galaxy, a disc of ionized and molecular gas (Boizelle et al. 2017; Ruffa et al. 2019) is observed, which is also visible in the SOAR image in the right-hand panel of Fig. 8. The circumnuclear dust disc seen by HST in the right-hand panel of Fig. 8, could be covering the central engine of the galaxy, offering a possible explanation for the apparent strongly sub-Eddington nuclear luminosity of L_bol/L_Edd⁴ ∼ 2 × 10⁻⁵ (Pellegrini et al. 2003). The true Eddington ratio necessary to explain the acceleration of jets penetrating through the ambient medium could be significantly higher. On the other hand, the low Eddington ratio could also be a result of an advection-dominated accretion flow, where the heat generated via viscous dissipation in the disc is ‘advected’ inwards with the accreting gas (Narayan & Yi 1994).

4.3 Peculiar bending jets in a hot atmosphere

The radio and X-ray morphologies of IC 4296 resemble those of some other systems, such as NGC 1265, where the powerful radio jets penetrate through the atmosphere of hot gas bound to the host galaxy, with little impact on the X-ray gas (Sun, Jerius & Jones 2005a). In NGC 1265, the two-sided jet emerges from the atmosphere into the ICM of the Perseus cluster. Since NGC 1265 is moving at high speed relative to the ICM, we see bright radio knots in its jet where they impinge on the ICM and are deflected (O’Dea & Owen 1987). Although IC 4296 differs in important respects from NGC 1265, we cannot exclude that the observed rebrightening in its jet at |$r\sim 10\, \rm {kpc}$| could have a similar origin.

The presence of a possible X-ray tail seen in the Chandra image and the location of IC 4296 in the centre of Abell 3565 is consistent with the sloshing of the ambient ICM in the potential well of the cluster (see Markevitch & Vikhlinin 2007). Given the relative motion, the radio jets emerging from the central atmosphere run into the ICM and are deflected in the direction opposite to the galaxy’s travel, as also noted by Kemp (1994). Assuming that the jets are supersonic, they could be deflected in a succession of inclined weak shocks or via magnetohydrodynamical instabilities (Massaglia et al. 2018), converting jets’ kinetic energy to thermal energy. Alternatively, the brightening could result from the dissipation of turbulence driven by the interaction between the jets and the external medium (Burns 1998). In either case, the additional thermal energy will raise the pressure (and density) in the jet plasma, causing it to expand (e.g. Gizani & Leahy 1999). This would account for the increase in the width of the radio source in the regions |${\sim }10\, \rm {kpc}$| from the central AGN, also noticed by Killeen & Bicknell (1988).

The apparent curvature of the jets is rather weak and we may be viewing the system from a direction nearly parallel to the velocity vector. The location of the tail and the radio source curvature are both consistent with a projected relative velocity towards the northeast.

5 SUMMARY AND CONCLUSIONS

We present the results of new VLA radio observations of the giant elliptical galaxy IC 4296, obtained in the A and D configurations. Combined with complementary X-ray and optical data, we provide a comprehensive, multiwavelength analysis of the AGN feedback, and jet–ICM interaction at play in this peculiar system. Our results can be summarized as follows:

• The new high-resolution VLA data in A configuration reveal the bright central region of IC4296 and two almost symmetrical knotty radio jets extending far beyond the host galaxy. With the VLA D configuration data, we confirm the presence of well-defined radio lobes first reported by Killeen et al. (1986b). The lobes, without significant hotspots, have large diameters of |${\sim }160\, \rm {kpc}$| at radii |$r\gtrsim 230\, \rm {kpc}$| and are undetected in the higher resolution A configuration VLA images.

• The XMM–Newton data reveal the presence of an X-ray cavity associated with one of the large radio lobes indicating a remarkable total enthalpy |${\sim }7\times 10^{59}\, \rm {erg}$|⁠. For comparison, this enthalpy is 30 times larger than for NGC 4261, due to the six times larger size of the radio lobes.

• The observed radio morphology is consistent with 3D hydrodynamical simulations (e.g. Massaglia et al. 2016) of supersonically expanding jets/lobes, with bow shocks created at the boundary layer of the radio lobes interacting with the ambient ICM.

• The jets are bent likely due to the relative motion of the ICM with respect to IC 4296 and magnetohydrodynamic instabilities, advecting the plasma in long narrow plumes. The relative motion of the galaxy with respect to the ICM is also supported by a possible X-ray tail seen in the Chandra image.

• The powerful jets piercing through the hot galactic atmosphere do not appear to be depositing sufficient heat in the innermost gas to prevent it from cooling, with the bulk of their energy deposited well beyond 10 kpc from the nucleus. The atmosphere of the galaxy can therefore continue to cool rapidly, feeding the central supermassive black hole and powering the jets. This is supported by the steep entropy and cooling time profiles of the hot galactic atmosphere which continue to drop all the way to the nucleus, where a warm Hα + [|$\rm{N\,{\small II}}$|] nebula and a cold molecular CO disc are clearly seen. As for a handful of other examples, such as NGC 4261, the cooling continues despite the powerful jet activity.

ACKNOWLEDGEMENTS

This work was supported by the Lendület LP2016-11 grant awarded by the Hungarian Academy of Sciences.

The NRAO is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. The scientific results reported in this article are based on observations made by the Chandra X-ray Observatory and published previously in cited articles.

Based on observations obtained at the SOAR telescope, which is a joint project of the Ministério da Ciência, Tecnologia, Inovações e Comunicações (MCTIC) do Brasil, the U.S. National Optical Astronomy Observatory (NOAO), the University of North Carolina at Chapel Hill (UNC), and Michigan State University (MSU). KG was supported by the János Bolyai Research Scholarship of the Hungarian Academy of Sciences.

This research has made use of resources provided by the Compagnia di San Paolo for the grant awarded on the BLENV project (S1618_L1_MASF_01) and by the Ministry of Education, Universities and Research for the grant MASF_FFABR_17_01.

This investigation is supported by the National Aeronautics and Space Administration (NASA) grants GO4-15096X, AR6-17012X, and GO6-17081X.

This work is supported by the ‘Departments of Excellence 2018–2022’ Grant awarded by the Italian Ministry of Education, University and Research (MIUR) (L. 232/2016).

FM acknowledges financial contribution from the agreement ASI-INAF no. 2017-14-H.0. ACF acknowledges support from ERC Advanced Grant 340442. KR acknowledges financial support from the ERC Starting Grant ‘MAGCOW’, no. 714196 and MS acknowledge the support from the NSF grant 1714764.

Footnotes

1

The GLEAM restoring beam is |$155.57\, \rm {arcsec}\times 143.65\, \rm {arcsec}$|⁠.

2

CTIO (Cerro Tololo Inter-American Observatory) H_α filter 6649/76 (λ_cen = 6650 Å, and FWHM  = 77 Å) and 6737/76 (λ_cen = 6746 Å, and FWHM  = 86 Å).

3

With corresponding flux density |${\sim }8\, \mathrm{ Jy}$|⁠.

4

The Eddington luminosity for IC 4296 is |${\sim }\, 1.2\times \, 10^{47}\, \rm {erg\, s^{-1}}$|⁠.