The nuclear activity and central structure of the elliptical galaxy NGC 5322

Abstract

We have analysed a new high-resolution e-MERLIN 1.5 GHz radio continuum map together with HST and SDSS imaging of NGC 5322, an elliptical galaxy hosting radio jets, aiming to understand the galaxy’s central structure and its connection to the nuclear activity. We decomposed the composite HST + SDSS surface brightness profile of the galaxy into an inner stellar disc, a spheroid, and an outer stellar halo. Past works showed that this embedded disc counter-rotates rapidly with respect to the spheroid. The HST images reveal an edge-on nuclear dust disc across the centre, aligned along the major-axis of the galaxy and nearly perpendicular to the radio jets. After careful masking of this dust disc, we find a central stellar mass deficit M_def in the spheroid, scoured by SMBH binaries with final mass M_BH such that M_def/M_BH ∼ 1.3–3.4. We propose a three-phase formation scenario for NGC 5322, where a few (2–7) ‘dry’ major mergers involving SMBHs built the spheroid with a depleted core. The cannibalism of a gas-rich satellite subsequently creates the faint counter-rotating disc and funnels gaseous material directly on to the AGN, powering the radio core with a brightness temperature of T_{B, core} ∼ 4.5 × 10⁷ K and the low-power radio jets (P_jets ∼ 7.04 × 10²⁰ W Hz⁻¹), which extend ∼1.6 kpc. The outer halo can later grow via minor mergers and the accretion of tidal debris. The low-luminosity AGN/jet-driven feedback may have quenched the late-time nuclear star formation promptly, which could otherwise have replenished the depleted core.

galaxies: elliptical and lenticular, cD, galaxies: nuclei, galaxies: photometry, galaxies: structure, radio continuum: galaxies

1 INTRODUCTION

Elliptical galaxies are thought to have formed hierarchically via mergers of smaller systems and accretion events (e.g. White & Rees 1978; Cole et al. 1994; Khochfar & Burkert 2001; Springel et al. 2005; Naab, Khochfar & Burkert 2006; Hopkins et al. 2009; Mundy et al. 2017). Merger-driven inflow of gas was invoked to produce nuclear starbursts or fuel the central black hole. The latter may drive powerful relativistic jets that can heat the surrounding gas, and as such impact on the central stellar structures in elliptical galaxies (e.g. Carlberg 1986; Hernquist 1992, 1993; Mihos & Hernquist 1994; Bahcall et al. 1995; Surace, Sanders & Evans 2001; Croton et al. 2006; Werner et al. 2014; Somerville & Davé 2015). Observations reveal that elliptical galaxies possess distinct structures on scales of a few kiloparsecs down to tens of parsecs such as depleted cores, nuclear star clusters, small-scale stellar discs, nuclear dust discs/lanes, and AGN (e.g. Crane et al. 1993; Kormendy et al. 1994; Jaffe et al. 1994; Ferrarese et al. 1994; Lauer et al. 1995; Carollo et al. 1997, Faber et al. 1997; Knapen, Shlosman & Peletier 2000; Rest et al. 2001; Trujillo et al. 2004; Lauer et al. 2005, 2007; Côté et al. 2006; Ferrarese et al. 2006; Kormendy et al. 2009; Dullo & Graham 2012, 2014; Turner et al. 2012; Savorgnan & Graham 2016; Dullo, Graham & Knapen 2017). However, the interplay between such central structures and the nuclear activity of the galaxies remain largely unclear.

Luminous elliptical galaxies brighter than M_B ∼ −20.5 ± 0.05 mag contain partially depleted cores, with sizes typically 20–500 pc, that are thought to be scoured by binary supermassive black holes formed in gas-poor (‘dry’) major mergers (e.g. Begelman, Blandford & Rees 1980; Ebisuzaki, Makino & Okumura 1991; Faber et al. 1997; Merritt 2006). The coexistence of such cores and rapidly rotating nuclear stellar and dust discs in galaxies is of interest since it may lead us to understand how such galaxies can acquire and transfer gaseous material inward to build the nuclear stellar/dust discs, fuel the central AGN, but not refill the depleted core with young stars. The radio elliptical (E3−4) galaxy NGC 5322 (RC3; de Vaucouleurs et al. 1991), with a depleted core, i.e. a central stellar mass deficit (e.g. Carollo et al. 1997; Richings, Uttley & Körding 2011; Krajnović et al. 2013a), nuclear stellar disc and LINER-type nuclear emission (Baldi & Capetti 2009), is a promising candidate which allows detailed, multiwavelength analysis of the central structure, AGN activities and jet structures.

The Legacy e-MERLIN Multi-band Imaging of Nearby Galaxies (LeMMINGs; Beswick et al. 2015) survey is a high-resolution radio continuum survey designed, in part, to study such galaxies by utilizing synergies from a large sample of high-resolution, multiwavelength data, including from HST: optical plus IR, Spitzer: IR, Herschel: IR, Chandra: X-ray and GALEX: UV. The project is composed of two complementary radio surveys: moderately deep snapshots of a large sample of 280 nearby (3–100 Mpc) galaxies over a large luminosity, size and morphological type range, selected from the Palomar Bright Galaxies Survey (Ho, Filippenko & Sargent 1997), and a deep survey of 6/280 representative galaxies. These radio surveys coupled with the ancillary data will probe the central structures, AGN/jet structure and star formation rates in galaxies (see Westcott et al. 2017; Williams et al. 2017). NGC 5322 is among the first galaxies observed with e-MERLIN at 1.5 GHz as part of the moderately deep LeMMINGs survey.

Isophotal analysis of ground-based optical images by Bender (1988) showed that the isophote shapes in NGC 5322 change from discy isophotes at R ≲ 8 arcsec, due to the galaxy’s small-scale, edge-on stellar disc, to boxy isophotes at larger radii (⁠|$8\,\text{arcsec} \lesssim R\lesssim 35\,\text{arcsec}$|⁠). Embedded discs are common in elliptical galaxies (e.g. Rest et al. 2001, references therein; Dullo & Graham 2012; Graham, Ciambur & Savorgnan 2016). The Bender (1988) kinematic data for NGC 5322, out to 35 arcsec, also revealed that the inner region |$(R \lesssim 2.6\,\text{arcsec})$|⁠, where the embedded stellar disc contributes light significantly, is counter-rotating with respect to the rest of the galaxy (see also Bender & Surma 1992, Rix & White 1992 and Krajnović et al. 2011, their fig. C4). This counter-rotating core (KDC) exhibits enhanced Mg₂ relative to the outer parts of the galaxy (Bender & Surma 1992; Kobayashi & Arimoto 1999). The Mg₂ line index is thought to trace primarily metallicity. The stellar population analysis within 1R_e by Spolaor et al. (2010, and references therein) has revealed that the galaxy is very metal rich ([Z/H] ∼0.45) with a super-solar α-abundance ratio ([α/Fe] ∼ 0.30) in the central region. Carollo et al. (1997) found that the V − I colour of the KDC differs only slightly ( ≲ 0.02 mag) from that of the surrounding regions. These isophotal, kinematical and stellar population properties and the colour of NGC 5322 were attributed to the build-up of the galaxy through a major merger of two gas-rich galaxies (e.g. Schweizer et al. 1990; Hernquist & Barnes 1991; Bender & Surma 1992; Schweizer & Seitzer 1992). However, the presence of a depleted core in NGC 5322 argues against such a gas-rich, major merger scenario in which the merger-driven gas dissipation and ensuing nuclear starburst are thought to produce coreless galaxies.

Reducing the Tonry et al. (2001) distance moduli by 0.06 mag (Blakeslee et al. 2002) yields a distance of 30.3 Mpc for NGC 5322. The galaxy has a central velocity dispersion σ ∼ 229 km s⁻¹ (HyperLeda,¹ Paturel et al. 2003). We assume that H₀ = 70 km s⁻¹ Mpc⁻¹ and that 1 arcsec corresponds to 145 pc.

The aim of this work is to investigate the formation mechanisms for NGC 5322 using a multiwavelength analysis of the galaxy’s central structure and the connection to the nuclear activity. To achieve this, we use our new 1.5 GHz e-MERLIN observations plus HST and SDSS imaging. These data and the pertaining data reduction steps are discussed in Section 2. After carefully masking the galaxy’s nuclear dust disc, for the first time, we decompose the 200 arcsec composite (HST + SDSS) light profile into an inner (exponential) stellar disc, a core-Sérsic spheroid and an outer exponential stellar halo (Section 3). Section 3 also details our radio, isophotal, and light profile fitting analyses, along with a literature comparison. Sections 4.1 and 4.2 provide our measurements of the central stellar mass deficit and merger rate for NGC 5322. In Section 4.3, we discuss the formation of NGC 5322. Section 5 summarizes our main conclusions.

2 OBSERVATIONS AND DATA REDUCTION

2.1 e-MERLIN radio data

High-sensitivity, high-resolution radio continuum observations of NGC 5322 were carried out using the e-MERLIN array at L band (1.25–1.75 GHz) on 2015 May 17 as part of the e-MERLIN LeMMINGs legacy programme (Beswick et al. 2014). We used all seven of the e-MERLIN telescopes throughout these observations, including the 76-m Lovell telescope. The target source was observed for a total on-source time of 5.7 h. These observations were regularly interspersed with short scans of the nearby phase calibration source, J1335+5844, with a 13.2 min target to phase calibrator cycle. Observations of flux density calibration source, 3C286, and bandpass calibrator, OQ208, were made at the beginning and end of the observing run. A total bandwidth of 512 MHz was correlated into eight adjacent spectral windows with each spectral window correlated into 512 frequency channels per polarization. Following initial data editing and the excision of RFI, these data were averaged in frequency (0.5 MHz) and time (2 s) before being processed using the e-MERLIN pipeline (Argo 2015) to apply standard multifrequency calibration procedures to determine and apply the complex gain solutions derived from the calibration sources. Following these pipeline procedures, a number of self-calibration cycles were carried out on the target source, NGC 5322, before reweighting of the relative sensitivities of the individual e-MERLIN telescopes (as a function of time and frequency) and final imaging.

Multifrequency synthesis imaging techniques were applied to the calibrated target source data using a variety of weighting schemes optimizing for sensitivity and angular resolution. The final images (see Fig. 1d) with a restoring beam of 0|${^{\prime\prime}_{.}}$|43 × 0|${^{\prime\prime}_{.}}$|17, an ∼41 × 41 arcsec² field of view (FOV), rms noise level of 11.6 μJy beam⁻¹ and peak target flux density of 5.26 mJy beam⁻¹ were used for subsequent analysis.

Figure 1.

(a) HST ACS F814W image of NGC 5322. (b) ACS F814W residual image of NGC 5322 created by subtracting the ellipse model image from the ACS F814W image. (c) HST ACS F435W − F814W colour map of the NGC 5322. The relatively redder (F435W − F814W ≳ 2.3) nuclear regions are due to the nuclear dust disc with radius R ∼ 1|${^{\prime\prime}_{.}}$|7 ≈ 246.5 pc and (d) e-MERLIN 1.5 GHz radio continuum map of NGC 5233. The beam is shown in the lower left corner. In all the panels north is up, and east is to the left. The nuclear dust disc is embedded in the small-scale stellar disc with a half-light radius R_e ∼ 356.7 pc. Both the dust and stellar discs are edge-on and aligned along the major axis of the galaxy, i.e. in the east–west direction. The dark regions, bright cones along the major-axis and the ‘X’-shaped structure in the residual image are due to the dust disc, stellar disc and boxy spheroid, respectively.

2.2 HST imaging data

HST Advanced Camera for Surveys (ACS; Ford et al. 1998) Wide Field Channel (WFC) F435W and F814W images of NGC 5322 (HST Proposal ID 9427;Harris et al. 2006) were retrieved from the public Hubble Legacy Archive (HLA²). The ACS F435W and F814W filters are similar to Johnson’s B and I bands, respectively. In addition, we retrieved the HST Wide Field Camera 3 (WFC3) IR F110W image from the Mikulski Archive for Space Telescopes (MAST³) to better account for the edge-on nuclear dust disc visible in the images and in our colour map (Figs 1–3). The ACS WFC and WFC3 IR cameras have pixel scales of 0|${^{\prime\prime}_{.}}$|05 pixel⁻¹ and 0|${^{\prime\prime}_{.}}$|13 pixel⁻¹, respectively.

Figure 2.

Top panels: HST ACS F814W image of NGC 5322 showing the central regions of the galaxy with the nuclear dust disc (left) and the corresponding masked regions (black areas) overplotted on the image (right). North is up and east is to the left. Bottom panels: as in the top panels but for the HST WFC3 F110W image. The orientation of the WFC3 images is marked with arrows.

Figure 3.

e-MERLIN 1.5 GHz contours of NGC 5322 overlaid on the (a) residual image and (b) F435W − F814W colour map of the galaxy, highlighting the residual structures and colour excess due to the dust disc. We overplot the elliptical isophotes derived from the ellipse fit to the ACS F814W image (see Section 2.4) to illustrate the position angle (PA) of the galaxy relative to the dust disc and radio jets. For clarity we do not show a few isophotes inside R ≲ 0|${^{\prime\prime}_{.}}$|4. While the black contour (with a semi-major axis of ∼1|${^{\prime\prime}_{.}}$|7) encloses the masked region shown in Fig. 2, we use the ACS F814W light profile at |$3\,\text{arcsec} \lesssim R \lesssim 40\,\text{arcsec}$| (Section 2.4). The radio jet is nearly perpendicular to the nuclear stellar and dust discs that are aligned along the major axis of the galaxy and result in the bright cones in east–west direction (a) and the colour excess (b), respectively. In both panels, north is up and east is to the left. The contour levels are (−1.00, 1.00, 1.41, 2.00, 2.83, 4., 5.66, 8.00, 11.31, 16.00, 22.63, 32.00, 45.25, 64.00, 90.51, 128.00, 181.00, 256.00, 362.00, and 512.00) × 3.326E-05 Jy beam⁻¹.

2.3 SDSS imaging data

We obtained the SDSS i-band mosaic image of NGC 5322, with a pixel scale of 0|${^{\prime\prime}_{.}}$|396 and a ∼ 18 × 18 arcmin⁻² FOV from the Data Release 13 (DR13⁴) data base. This image was used to determine reliably the sky background of NGC 5322.

2.4 Surface brightness profile

We follow the reduction procedure by Dullo, Martínez-Lombilla & Knapen (2016, see references therein) and fit elliptical isophotes to the WFC3 F110W, ACS F814W and SDSS i-band images of NGC 5322 using the iraf task ellipse (Jedrzejewski 1987) to extract the surface brightness, ellipticity (ε = 1 − b/a, where a and b are the semimajor and minor-axes of the best-fitting ellipse), PA, and isophote shape parameter (B₄) profiles along the major-axis (see Fig. 3b). B₄ quantifies deviations of the isophotes from pure ellipses (i.e. B₄ = 0). If B₄ < 0 then the isophotes are boxy, B₄ > 0 if the isophotes are discy. We run ellipse on the WFC3 F110W, ACS F814W, and SDSS i-band images using a mask for each image that is generated by SExtractor (Bertin & Arnouts 1996) together with a manual mask to exclude the galaxy’s nuclear dust disc, gaps between CCD detectors in the images, background galaxies, bright foreground stars, and chip defects.

Fig. 4 shows a composite light profile of NGC 5322 together with the pertaining ε, PA and B₄ profiles, assembled by combining our WFC3 F110W, ACS F814W, and SDSS i-band data. At R ≲ 3 arcsec, we used the near-IR WFC3 F110W light profile which is significantly less contaminated by dust extinction (see Sections 3.2 and 4.2.1). In Fig. 2, we show the central regions and associated masks of NGC 5322 to highlight the masked nuclear dust regions in the WFC3 F110W and ACS F814W images. The ACS F814W data, with a better spatial resolution than the WFC3 and SDSS data, were used at |$3\,\text{arcsec} \lesssim R \lesssim 40\,\text{arcsec}$|⁠. In order to capture NGC 5322’s stellar halo that extends beyond the FOVs of the ACS and WFC3 images, at large radii (R ≳ 40 arcsec), we used the SDSS i-band data to constrain the sky-background. The WFC3 F110W and SDSS i-band data were shifted to match the ACS F814W data. Although the WFC3 F110W filter covers a wavelength range mostly redder than those of the ACS F814W- and SDSS i-band filters, we find a good overlap between the ACS F814W, SDSS i, and WFC3 F110W light profiles over the R ≲ 10 arcsec radial range, when we exclude the dust-affected region. The magnitudes in this paper are in the VEGA magnitude system.

Figure 4.

Three-component inner disc (dash–dotted green curve) + boxy spheroid (dotted magenta curve) + outer halo (dashed blue curve) model fit to the composite HST WFC3 IR F110W + HST ACS F814W + SDSS i-band surface brightness profile of NGC 5322 (see Section 2.4). The solid black curve shows the complete fit to the profile. The WFC3 F110W and SDSS i-band light profiles were shifted to match the ACS F814W data. This decomposition yields a depleted core with a break radius of R_b ∼ 0|${^{\prime\prime}_{.}}$|37 ± 0|${^{\prime\prime}_{.}}$|04. Note that we extracted the innermost (R ≲ 3 arcsec) light profile of the galaxy using (i) the WFC3 IR F110W image and (ii) a careful mask to avoid the nuclear dust disc, with a thickness of ∼0|${^{\prime\prime}_{.}}$|2 ≈ 28 pc near the centre of the galaxy, dividing the galaxy centre (Fig. 2).

3 STRUCTURAL ANALYSIS

3.1 Residual image and dust map

Fig. 1(b) shows the residual image of NGC 5322 generated by subtracting the iraf ellipse model image from the ACS F814W image. Four structural features are evidenced in this image: (i) a dark nuclear disc along the galaxy’s major axis which is due to the edge-on inner dust disc (with a radius of 1|${^{\prime\prime}_{.}}$|7 ≈ 246.5 pc and a thickness of 0|${^{\prime\prime}_{.}}$|6 ≈ 87 pc near the galaxy centre, respectively) that is visible in the galaxy images crossing the centre of the galaxy, (ii) two bright cones along the major axis, (iii) two bright cones along the minor-axis, and (iv) a large-scale ‘X’-shaped structure due to the spheroid’s boxy isophotes (see also Fig. 3a and Section 3.4). The dust disc, with F435W − F814W ∼ 2.72|$^{+0.13}_{-0.39}$| compared to the relatively bluer immediate surroundings (F435W − F814W ∼ 2.30), is better visible in the colour maps (Figs 1c and 3b), rather than in the residual image, as an inner disc with red colour excess of ∼0.42 mag (see also Carollo et al. 1997; Tomita et al. 2000). To create this colour map, we registered the F435W and F814W images using the iraf tasks geomap and gregister. Excluding the dust disc, the galaxy becomes progressively bluer towards larger radii, with an F435W − F814W colour gradient of ∼0.22 mag per 10 arcsec. We note that the dust disc has a thickness of ∼0|${^{\prime\prime}_{.}}$|2 ≈ 28 pc near the galaxy centre in the WFC3 F110W image (Fig. 2).

3.2 Structural decomposition

NGC 5322 is well known for its inner (discy) stellar component, which counter-rotates with respect to the galaxy’s boxy spheroid (e.g. Bender 1988, 1996; Bender & Surma 1992; Rix & White 1992; Goudfrooij et al. 1994; Scorza & Bender 1995; Carollo et al. 1997; Silva & Bothun 1998), and for its central radio source (Feretti et al. 1984; Hummel, van der Hulst & Dickey 1984). Analysis of high-resolution HST ACS/WFC images shows that the galaxy contains a small-scale discy stellar component, a boxy spheroid (that has a depleted core as revealed by the flattened inner stellar distribution which deviates downward relative to the inward extrapolation of the spheroid’s outer Sérsic profile) and an outer stellar halo (Figs 1, 3, and 4). As mentioned in the Introduction, early-type galaxies brighter than M_B ∼ −20.5 mag, hereafter core-Sérsic galaxies, have depleted cores and central stellar mass deficits which are thought to be created through the ejection of stars by inspiralling binary SMBHs (not due to central dust obscuration), while those with |$M_{\text{B}} \gtrsim -20.5$| mag tend to be coreless (e.g. Faber et al. 1997; Kormendy 1999; Ravindranath et al. 2001; Rest et al. 2001; Trujillo et al. 2004; Lauer et al. 2005, 2007; Ferrarese et al. 2006; Dullo & Graham 2012, 2013, 2014; Rusli et al. 2013).

The light distributions of core-Sérsic galaxies are well described by the core-Sérsic model⁵ which is a blend of an inner power law and an outer Sérsic model (Graham et al. 2003). Detailed multicomponent decompositions using high-resolution light profiles, extending to at least 1R_e (Graham et al. 2003), are crucial to understand properly the different formation mechanisms that build up the distinct structural components in galaxies. Therefore, we fit a three-component Sérsic small-scale disc + core-Sérsic spheroid + exponential halo model to the major-axis surface brightness profile of NGC 5322 which extends out 200 arcsec. Fig. 4 shows this light profile decomposition with a small rms residual scatter of ∼0.015 mag arcsec⁻². Each model component was convolved with a Gaussian PSF with FWHM ∼0|${^{\prime\prime}_{.}}$|25 (see Dullo et al. 2017, their section 3.1). The FWHM of the PSF was determined using several stars in the WFC3 F110W image of the galaxy. Table 1 provides the best-fitting model parameters.

Table 1.

Structural parameters for NGC 5322 resulting from the fit to the optical data.

Galaxy	Type	μb	Rb	Rb	γ	α	ncS	Re, cS	Re, cS	μe, S	nS	Re, S	Re, S	μ0, halo	h	h
			(arcsec)	(pc)				(arcsec)	(kpc)			(arcsec)	(pc)		(arcsec)	(kpc)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)
NGC 5322	E3-4	14.36	0.37	53.7	0.17	2	4.0	30.4	4.4	17.21	1.0	2.46	356.7	20.79	53.6	7.8
			10 per cent		10 per cent		20 per cent	25 per cent			20 per cent	20 per cent			10 per cent

Notes. – First row: structural parameters from the three-component Sérsic inner disc + core-Sérsic spheroid + exponential halo fit to the F814W-band (∼I-band) surface brightness profiles of NGC 5322 (Fig. 4). Col. 1: galaxy name. Col. 2: morphological type. Cols. 3−10: best-fitting parameters from the core-Sérsic model. Cols. 11 − 14: Sérsic model parameters for the exponential (n = 1) inner stellar disc with scale length (h = R_{e, S}/1.678) ∼ 1|${^{\prime\prime}_{.}}$|47 and central surface brightness (μ_{0, disc}) ∼ μ_{e, S} − 1.822 ∼ 15.39 mag arcsec⁻². Cols. 15−17 best-fitting parameters for the exponential halo light. μ_b, μ_e and μ_{0, halo} are in mag arcsec⁻². The second row shows the uncertainties on the fit parameters. The uncertainty associated with μ_b, μ_{e, S} and μ_{0, h} is ∼ 0.02 mag arcsec⁻².

We found that a Sérsic model fit to the halo light yields n ∼ 1.0 ± 0.2. We therefore force an exponential (n = 1) model for the outer stellar halo. Similarly, the embedded, small-scale discy component was modelled using a Sérsic model with a half-light effective radius R_e ∼ 2|${^{\prime\prime}_{.}}$|46 ± 0|${^{\prime\prime}_{.}}$|49 ≈ 356.7 ± 71 pc and n ∼ 1.0, indicating an inner exponential disc with a scale length h ∼ 1|${^{\prime\prime}_{.}}$|47 ± 0|${^{\prime\prime}_{.}}$|15 ≈ 212.6 ± 21.3 pc and an I-band central surface brightness μ_{0, disc} ∼ 15.39 ± 0.02 mag arcsec⁻². The boxy spheroid which dominates over R ≲ 100 arcsec has a core-Sérsic stellar light distribution with n ∼ 4.0 ± 0.8 (i.e. essentially the de Vaucouleurs (1948) R^1/4 law outside the depleted core), R_b ∼ 0|${^{\prime\prime}_{.}}$|37 ± 0|${^{\prime\prime}_{.}}$|04, μ_{b, I} ∼ 14.36 ± 0.02 mag arcsec⁻², and R_e ∼ 30|${^{\prime\prime}_{.}}$|4 ± 7|${^{\prime\prime}_{.}}$|6. At R ≳ 100 arcsec, the spheroid and outer stellar halo contribute equally.

3.3 Literature comparison

For comparison, Rix & White (1992) decomposed the radially limited ∼10″ absorption line profile of NGC 5322, which does not probe the outer parts of this galaxy’s spheroid and stellar halo, into an inner stellar disc with h ∼ 2|${^{\prime\prime}_{.}}$|1 and a spheroid with R_e ∼ 25 arcsecs using a double Gaussian model. Direct comparison between our decomposition and that by Rix & White (1992) is challenging given the different approaches used by these two works, i.e. light profile versus absorption line profile decompositions. Scorza & Bender (1995, their fig. A.9) fit an exponential disc + R^1/4 bulge model to their ∼10 arcsec data with a seeing FWHM of 1|${^{\prime\prime}_{.}}$|6 and a plate scale of 0|${^{\prime\prime}_{.}}$|46/pixel, finding R_e ∼ 3.4 kpc for the bulge, and h ∼ 180 pc, a V-band μ_{0, disc} ∼ 19.45 mag arcsec⁻².

Fig. 5 compares our break radius (R_b) and inner logarithmic slope of the spheroid’s inner light profile (−γ) with those from Carollo et al. (1997), Richings et al. (2011), and Krajnović et al. (2013a). Carollo et al. (1997) and Krajnović et al. (2013a) fit the Nuker model (Lauer et al. 1995, 2005) to 10 arcsec HST WFPC2/PC1 light profile of NGC 5322, while Richings et al. (2011) modelled the galaxy’s HST ACS/WFC light profile using the core-Sérsic model. Also, Krajnović et al. (2013b, their fig. 1) modelled their ground-based light profile of the galaxy over the radius range 2|${^{\prime\prime}_{.}}$|5 ≲ R ≲ 172 arcsec using a single Sérsic model, resulting in a residual profile which clearly reveals structures that have not been properly modelled. Because these past works (Carollo et al. 1997; Richings et al. 2011; Krajnović et al. 2013b, 2013a) treated the galaxy as a single-component system, the fraction of the light attributed to the spheroid in their modellings come from the inner stellar disc and outer halo. This led, in part, to the measurements of the break radius/half-light radius and Sérsic index by Richings et al. (2011, R_b ∼ 1|${^{\prime\prime}_{.}}$|23, n ∼ 8.1 and R_e ∼ 93|${^{\prime\prime}_{.}}$|1, their table 2) and Krajnović et al. (2013b, n ∼ 5.8, R_e ∼ 67|${^{\prime\prime}_{.}}$|9, their fig. 1) to be biased high (Fig. 5). Measuring the core size as the radius where the negative logarithmic slope of the fitted Nuker model profile equals 1/2 (R_γ = 1/2), Krajnović et al. (2013a, their table C1) found R_γ = 1/2 ∼ 0|${^{\prime\prime}_{.}}$|71, while Carollo et al. (1997, their table 4) reported R_γ = 1/2 ∼ 0|${^{\prime\prime}_{.}}$|38, which agrees with our core-Sérsic break radius for the galaxy (Fig. 5a).

Figure 5.

Comparison of our break radius (R_b) and inner logarithmic slope of the spheroid’s inner light profile (−γ) with those values from Carollo et al. (1997), Richings et al. (2011), and Krajnović et al. (2013a). The solid lines show a one-to-one relation.

Part of the discrepancies between our work and Carollo et al. (1997), Krajnović et al. (2013a) and Richings et al. (2011) arises from the treatment of the inner dust disc. Indeed, the inner negative logarithmic slopes presented in Carollo et al. (1997, γ = 0.0) and Krajnović et al. (2013a, γ = 0.05) are shallower than that of ours (γ ∼ 0.17 ± 0.02), while Richings et al. (2011) reported a very steep inner negative logarithmic slope for the galaxy (γ = 0.41, Fig. 5b). Furthermore, replacing the inner R ≲ 3 arcsec profile in Fig. 4, extracted from the redder WFC3 F110W image, with the ACS F814W data and performing a three-component decomposition of the ACS F814W (R ≲ 40 arcsec) + SDSS i-band (R > 40 arcsec) profile yields a core with R_b ∼ 0|${^{\prime\prime}_{.}}$|49 ± 0|${^{\prime\prime}_{.}}$|05 ≈ 71.0 ± 7.3 pc, compared to that adopted here (R_b ∼ 0|${^{\prime\prime}_{.}}$|37 ± 0|${^{\prime\prime}_{.}}$|04 ≈ 53.7 ± 5.4 pc, see Fig. 4 and Table 1). In Section 4.2.1, we measure low F110W-band extinction (A_F110W ∼ 0.031 − 0.047 mag) in the regions immediately surrounding the masked regions near the galaxy centre (Figs 2 and 3). Also, we show that our break radius for this galaxy is in good agreement with those predicted based on the Dullo & Graham (2014) R_b–σ and R_b–M_V relations, i.e. |$R_{\rm b, \sigma \text{-}based} \approx 44.0 \pm 6.1$| pc and |$R_{{\rm b},M_{V}\text{-}{\rm based}} \approx 48.6 \pm 7.2$| pc, respectively (Section 4.2). This R_b agreement, together with the low-IR extinction, gives us some confidence that our measurement of NGC 5322’s break radius was not affected by the nuclear dust disc.

3.4 Ellipticity, position angle, isophote shape, and kinematics

Inside R ∼ 0|${^{\prime\prime}_{.}}$|3, NGC 5322 shows radial variations in ellipticity, PA, and B₄ due to the nuclear dust disc and the PSF (Figs 1 and 4). Going outward from R ∼ 0|${^{\prime\prime}_{.}}$|3, the ellipticity continuously rises from ε ∼ 0.10 to a maximum of ε ∼ 0.45 at ∼2|${^{\prime\prime}_{.}}$|5 and it then declines reaching a local minimum of ∼0.25 at ∼10 arcsec, before starting to gently rise again. The ellipticity profile, together with the positive B₄ values over 1|${^{\prime\prime}_{.}}$|5 − 8″, reveals the small-scale disc, consistent with the three-component profile decomposition (Fig. 4). The spheroid of NGC 5322 contains boxy isophotes, with ellipticity ε ∼ 0.35 ± 0.07, that resulted in the ‘X’-shaped structure in the residual image (Figs 1 and 3). Bender (1988, see also Rix & White 1992; Scorza & Bender 1995 and Krajnović et al. 2013b) found that the inner region (R ≲ 2|${^{\prime\prime}_{.}}$|6), where this galaxy’s inner disc contributes significantly, counter-rotates with maximum rotational velocity V_{rot, max} ∼ 80 km s⁻¹ relative to the intermediate parts the spheroid at |$10 \text{arcsec} \lesssim R \lesssim 35$| arcsec (V_{rot, max} ∼ −30 km s⁻¹). Rix & White (1992, their fig. 9) reported V_{rot, max}/σ ∼ 1.4 and ∼0.15 for the inner disc and spheroid, respectively.

The small-scale stellar disc, spheroid, and outer stellar halo of NGC 5322 are well aligned as revealed by the PA which remains stable at ∼94° outside the region most affected by the PSF (see Fig. 4).

3.5 Radio structure

NGC 5322 is a radio source which hosts a LINER nucleus (Baldi & Capetti 2009). Fig. 1(d) shows the 1.5 GHz e-MERLIN radio continuum map of NGC 5322 with a restoring beam of θ_maj × θ_min ∼ 0|${^{\prime\prime}_{.}}$|43 × 0|${^{\prime\prime}_{.}}$|17 at a PA ∼ 0.79° (Table 2). Lower resolution images of this source in the Faint Images of the Radio Sky at Twenty Centimetres (FIRST) survey (Becker, White & Helfand 1995) at 1.4 GHz show that it has an extended radio jet structure along the same PA of the inner jet structure observed in Fig 1. Whilst the VLA maps of the galaxy by Hummel et al. (1984) also revealed that the galaxy has an FRI radio morphology with twin radio jets extending over ∼20 arcsec at 1.4 GHz. These results are consistent but substantially shallower and at lower angular resolution, as compared to these new e-MERLIN observations. While the large bandwidth of these observations provides near complete aperture coverage to emission on scales between 0.1 and several arcsec, it should be noted that the shortest project baseline of e-MERLIN (∼11 km) in these observations means that these data are insensitive to very extended, diffuse emission. In the case of NGC 5322, this spatial filtering acts to separate the central jet emission from diffuse radio emission. Fig. 1(d) shows that these radio jets extend ∼10|${^{\prime\prime}_{.}}$|9 ≈ 1.6 kpc at a PA of 172° as observed with e-MERLIN. On large scales, the radio jets show slight deviations with respect to the orientation of the inner jet.

Table 2.

e-MERLIN observation.

Frequency	θmaj × θmin	Beam PA	θmaj × θmin	Beam PA	Speak	Sint, core	TB, core	Pcore	Sint, jets	Pjets
(GHz)		(°)		(°)	(mJy beam−1)	(mJy)	(K)	(W Hz−1)	(mJy)	(W Hz−1)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
1.5	0|${^{\prime\prime}_{.}}$|43 × 0|${^{\prime\prime}_{.}}$|17	0.79	0|${^{\prime\prime}_{.}}$|20 × 0|${^{\prime\prime}_{.}}$|04	169.6 ± 0.094	5.38 ± 0.01	5.96 ± 0.02	4.5 × 107	6.55 × 1020	6.41 ± 0.02	7.04 × 1020

Notes. Col. 1: centrally weighted (i.e. nominal) frequency. Col. 2: synthesized beam major-axis and minor-axis dimensions. Col. 3: synthesized beam PA. Cols. 4 and 5: deconvolved beam dimensions and PA obtained from a single Gaussian component fit to the central radio emission using the aips task jmfit. Col. 6: jmfit peak flux density of the core emission. Col. 7: jmfit integrated flux density of the core emission. Cols. 8 and 9: brightness temperature and power of the radio core. Col. 10: integrated flux density of the jets. Col. 11: radio power of the jets. The quoted errors on the PA, S_peak, and S_{int, core} (Cols. 5–7) are those reported by the jmfit. The error on S_{int, jets} is that given by tvstat.

Fig. 3 shows the radio contours overlaid on the HST ACS residual image and the ACS F435W − F814W colour map. In Fig. 3(b), we overplot the elliptical isophotes obtained from the ellipse fit to the ACS F814W image to illustrate the PA of the galaxy relative to the nuclear dust disc and radio jets. Because of NGC 5322’s nuclear dust disc crossing the galaxy centre, we did not attempt to align these images by shifting the radio core to the central maximum in the ACS images. We determine that the offset between the optical and radio images can be up to ∼0|${^{\prime\prime}_{.}}$|3. The inner jet is nearly perpendicular to the nuclear dust and stellar discs that are aligned along the major axis of the galaxy (i.e. PA ∼ 94°). The radio jets do not have strong enough optical/near-IR emissions to be seen from the optical and near-IR images, or as excess optical/IR emission above the (inner disc + spheroid + halo) model fit to the galaxy light profile (Fig. 4).

3.5.1 Flux density, brightness temperature, and radio power

We fit a single elliptical Gaussian model to the central radio source using the aips⁶ task jmfit, deriving the position, peak and integrated intensity flux density of the core emission. Also, the jmfit fit yields deconvolved beam dimensions and PA after deconvolving the clean beam from the fitted component size (see Table 2).

The measurement of the brightness temperature of a radio source is a diagnostic tool that discriminates between the thermal free–free emission from H ii regions and non-thermal synchrotron radio continuum emission. The former has a nebular brightness temperature of T_B ∼ 10⁴ K, while the latter has a higher brightness temperature |$T_{\rm B} \gtrsim 10^{6}-10^{11}$| K (e.g, Condon 1992; Ho & Ulvestad 2001).

Using the Rayleigh–Jeans Law, the brightness temperature (T_B) can be written as

\begin{eqnarray} T_{\rm B} = \left(\frac{S_{\nu }}{\Omega _{\rm beam}\nu ^{2}} \right) \left(\frac{\text{c}^{2}}{2\text{k}}\right), \end{eqnarray}

(1)

where ν is the central frequency, k is the Boltzmann constant, c is the speed of light, Ω_beam is the deconvolved beam solid angle, and S_ν is the integrated radio flux density. For the radio core of NGC 5322, we derived T_B ∼ 4.5 × 10⁷ K at ν ∼ 1524 MHz, suggesting a non-thermal radio continuum emission. This confirms the conclusion from VLA 15 GHz and VLBA observations presented by Nagar, Falcke & Wilson (2005), who find a brightness temperature of >10^8.2 K.

We run the aips task tvstat to determine the integrated radio flux density for the jets S_{int, jets} ∼ 6.41 ± 0.02 mJy, which yields a 1.5 GHz radio power of ∼7.04 × 10²⁰ W Hz⁻¹ at NGC 5322’s distance of 30.3 Mpc (Table 2). Similarly, the core radio continuum emission of S_{int, core} ∼ 5.96 ± 0.02 mJy for NGC 5322 implies a core radio power of ∼6.55 × 10²⁰ W Hz⁻¹. The NRAO VLA Sky Survey (NVSS; Condon et al. 1998), with a very poor resolution of 45 arcsec, has detected a source, near NGC 5322’s centre, with a 1.4 GHz integrated flux of 78 ± 2.8 mJy. The FIRST survey finds three sources with integrated radio flux densities of ∼40.7, 8.8, and 8.1 mJy at ∼3|${^{\prime\prime}_{.}}$|1 (south), 7|${^{\prime\prime}_{.}}$|0 (north), and 11|${^{\prime\prime}_{.}}$|9 (north) from the galaxy centre, respectively. However, given the FIRST images’ resolution of 5 arcsec, this survey’s detection of the three sources and the pertaining flux values for NGC 5322 are not reliable. Nonetheless, our total recovered flux density (S_{int, jets} + S_{int, core} ∼ 12.4 mJy) is significantly lower (∼factor of 5) than that recovered for more compact/filled interferometer observations (e.g. FIRST, S_{int, total} ∼ 57.6 mJy) due to primarily the limited number of short baselines in the e-MERLIN array which makes it insensitive to structures greater than 5 arcsec. As such our estimate of the jet power for the galaxy is a lower limit.

4 DISCUSSION

4.1 Central mass deficit

The central stellar mass deficits of core-Sérsic galaxies are thought to be created by inspiralling binary SMBHs that form from ‘dry’ galaxy mergers (see Section 3.2). Theory predicts that the impact of multiple dry major mergers on core creation is cumulative. N-body simulations by Merritt (2006) suggested that the stellar mass deficit M_def that is generated by binary SMBHs after |$\mathcal {N}$| successive mergers equals |$0.5 \mathcal {N} M_{\rm BH}$|⁠, where M_BH is the total sum of the masses of the binary SMBHs. A careful measurement of the central stellar mass deficit in the spheroid of NGC 5322 can be used to constrain the galaxy’s merger history. We follow the procedures by Dullo & Graham (2014, their section 5.2) and compute the stellar luminosity deficit (L_def) as the difference in luminosity between the inwardly extrapolated outer Sérsic profile of the complete core-Sérsic model (fit to the spheroid’s profile) and the core-Sérsic model (see Section 3.2 and Fig. 4). This yields stellar luminosity deficit in the I band,⁷L_def/L_{⊙, I} ∼ 3.9 × 10⁸. To convert this luminosity deficit into M_def, we determined the stellar mass-to-luminosity ratio (M/L) excluding the inner dust disc and using the central dust-free colour of the galaxy (F435W − F814W ∼ 2.15, Fig. 1), which corresponds to V − I ∼ 1.24 (Fukugita, Shimasaku & Ichikawa 1995). For V − I ∼ 1.24, the colour–age–metallicity–(M/L) diagram by Graham & Spitler (2009, their fig. A1) yields an M/L ∼ 4.2 in the V band, which corresponds to an I-band M/L ∼ 2.8 (Worthey 1994, his table 5A), implying M_def ∼ (1.1 ± 0.3) × 10⁹ M_⊙.

4.2 Merger rate

Dynamical SMBH measurements are not available for NGC 5322. We therefore compute the M_def/M_BH ratios using SMBH masses predicted from the M_BH–σ and M_BH–L relations. For the velocity dispersion of NGC 5322 (σ ∼ 229 km s⁻¹), the Graham & Scott (2013) M_BH–σ relation predicts log (M_BH/M_⊙) ∼ 8.51 ± 0.40. To estimate the total I-band absolute magnitude of the the spheroid (M_{sph, I}), we integrate the core-Sérsic profile (Fig. 4) by taking into account the spheroid’s average ellipticity (ε ∼ 0.35). This yields M_{sph, I} ∼ −23.04 mag, corrected for Galactic extinction (∼− 0.02 mag) and surface brightness dimming (∼−0.03 mag), which corresponds to a V-band magnitude M_{sph, V} ∼ −21.77 mag applying the colour transformation V − I ∼ 1.27 (Fukugita et al. 1995). Converting the Graham & Scott (2013) B-band core-Sérsic M_BH–L relation into a V-band relation using B − V = 0.95 (Fukugita et al. 1995), we predict log (M_BH/M_⊙) ∼ 8.93 ± 0.34 for the spheroid of NGC 5322.

We thus find that M_def/M_BH ∼ 3.4 and 1.3 when using SMBH masses predicted based on the spheroid’s σ and L, respectively. While these numbers suggest that the spheroid has experienced |$\mathcal {N} \sim 2-7$| major mergers (Merritt 2006), this conclusion is hinged on the assumption that SMBH binary scouring is the sole mechanism that created the core. For comparison, past works have found mass deficits that are typically ∼0.5–4M_BH (Graham 2004; Trujillo et al. 2004; Ferrarese et al. 2006; Dullo & Graham 2013, 2014; Rusli et al. 2013). If the central mass deficit of NGC 5322 has been enlarged by additional mechanisms (such as the scouring action of gravitationally recoiled SMBH that oscillates about the centre of a merger remnant, e.g. Boylan-Kolchin, Ma & Quataert 2004; Gualandris & Merritt 2008, and the action of multiple SMBHs that form in galaxies at high redshifts, e.g. Kulkarni & Loeb 2012, see the discussion by Dullo & Graham 2014, their sections 5.3 and 5.4), then |$\mathcal {N} \lesssim 2-7$|⁠.

4.2.1 Dust absorption

Nuclear dust absorption can reduce the surface brightness of NGC 5322 at the inner regions, creating an artificial depleted core or enlarging the break radius/central mass deficit and resulting in |$\mathcal {N} \lesssim 2-7$| for the galaxy. However, not only did we carefully mask the inner dust disc to extract the WFC3 IR F110W profile at R ≲ 3 arcsec (see Sections 2.4, 3.1 and Fig. 2) but also the values of σ ∼ 229 km s⁻¹ and M_V ∼ −21.77 mag for NGC 5322’s spheroid imply that the Dullo & Graham (2014, their table 3) R_b–σ and R_b–M_V relations predict break radii of |$R_{\rm b, \sigma \text{-}based} \approx 44.0$| pc and |$R_{{\rm b},M_{V}\text{-}{\rm based}} \approx 48.6$| pc, in good agreement with the break radius from our modelling (Fig. 4, Table 1).

Below, as a further check on the effect of the dust absorption, we follow the dust correction steps described in Nowak et al. (2008, their appendix A, see also Bender et al. 2015, their section 3.3) and measure the F110W (∼J-band) extinction (A_J) in the nuclear regions of the galaxy using our F435W − F814W (B − I) colour map (Figs 1 and 3).

Assuming that dust acts as a foreground screen obscuring the stars, the extinction in the I band can be written as

\begin{eqnarray} A_{I}= I_{\text{obs}} - I_{\text{corr}} = \alpha _{\rm red} E(B-I), \end{eqnarray}

(2)

where I_obs and I_corr are the observed and extinction-corrected I-band surface magnitudes. The reddening E(B − I) = (B − I)_obs − (B − I)_corr and α_red = ((A_B/A_I) − 1)⁻¹. Therefore,

\begin{eqnarray} f_{I,\text{corr}}= f_{I,\text{obs}} \bigg (\frac{f_{I,\text{obs}}}{f_{B,\text{obs}}}\bigg )^{\alpha _{\rm red}} \bigg (\frac{f_{B,\text{corr}}}{f_{I,\text{corr}}}\bigg ) ^{\alpha _{\rm red}}, \end{eqnarray}

(3)

where f_I and f_B are the I- and B-band surface fluxes.

Assuming a very shallow stellar population gradient in the central region of NGC 5322 (Spolaor et al. 2010), we found the dust free surface flux ratio in the inner regions to be f_{B, corr}/f_{I, corr} ∼ 0.125. Following Schlafly & Finkbeiner (2011, their table 6), A_{ACS, F475W} = 3.610E(B − V), A_{ACS, F814W} = 1.526E(B − V) and A_{WFC3, F110W} = 0.881E(B − V) for R_V = 3.1. Thus, A_{ACS, F475W}/A_{ACS, 814W} ∼ 2.366, i.e. α_red ∼ 0.732 and A_{WFC3, F110W}/A_{ACS, 814W} ∼ 0.5773. Therefore, combining equa-tions (2) and (3) gives an I-band extinction value of A_I ∼ 0.080 mag for the regions, near the galaxy centre, immediately outside the ACS and WFC3 masked regions (Fig. 2) with a B − I colour of ∼2.37 mag. This A_I value corresponds to a low-IR extinction value of A_J ∼ 0.047 mag. For comparison, A_I ∼ 0.053 mag for the outer part of the dust disc (as traced by the colour map) near the galaxy centre with a B − I colour of ∼2.33 mag. For the reddest part of the dust disc with B − I ∼ 2.85 mag, which we have masked, A_I ∼ 0.053 mag (see Figs 1–3). Thus, we find a low-IR extinction value of A_J ∼ 0.031 mag for the outer part of the dust disc, while for the reddest part A_J ∼ 0.250 mag. These three A_J values can be compared to our adopted uncertainty on surface brightness μ_b, μ_e, and μ₀, i.e. ∼0.02 mag arcsec⁻².

We did not attempt to derive a dust-corrected F110W surface brightness profile for NGC 5322 by creating a dust-corrected WFC3 F110W image of the galaxy using equation (3). Although the foreground screen approximation (equation 2) enables us to constrain the uncertainties due to the effects of the nuclear dust, the technique is not fully reliable to derive a robust light profile for the galaxy (Nowak et al. 2008; Bender et al. 2015). Both Nowak et al. (2008, Appendix A) and Bender et al. (2015, Section 3.3) have estimated the best α_red values in a subjective manner by iterating α_red to obtain what they considered the best dust-corrected images for their galaxies.

4.3 Spheroid, inner disc, and outer halo formation and the connection to the central radio source

Here, we discuss the formation of NGC 5322, an E3 − 4 giant elliptical in a poor group (Garcia 1993; Madore, Freedman & Bothun 2004), considering the implications of our findings; a galaxy spheroid with a stellar mass deficit M_def ∼ (1.1 ± 0.3) × 10⁹ M_⊙, small-scale stellar and dust discs, an outer stellar halo and a central radio source (Section 3.5, see also Feretti et al. 1984; Hummel et al. 1984).

4.3.1 Spheroid property

In the previous sections, we have discussed that the depleted core in NGC 5322 is not due to the nuclear dust disc crossing the galaxy centre, which can create an artificial core or enlarge a pre-existing depleted core. However, it is still worth comparing the properties of the spheroid of NGC 5322 with those of core-Sérsic galaxies to ensure that our core identification is real. Faber et al. (1997) noted the connection between the ‘core’ versus ‘power law’ central structural dichotomy and the ‘boxy’ versus ‘discy’ and ‘slow rotator’ versus ‘fast rotator’ divides (see also Davies et al. 1983; Carter 1987; Bender 1988; Bender, Doebereiner & Moellenhoff 1988; Peletier et al. 1990; Jaffe et al. 1994). In agreement with this picture,⁸ NGC 5322’s core-Sérsic spheroid rotates slowly (e.g. Bender 1988; Rix & White 1992; Emsellem et al. 2011; Krajnović et al. 2013a,b) and has boxy isophotes (Figs 2 and 3). The spheroid’s Sérsic (n ∼ 4) stellar light distribution outside the core, M_V ∼ −21.77 mag ≈ 73.4 per cent of the total galaxy light (i.e. including the stellar halo) and σ ∼ 229 km s⁻¹, also suggest that the galaxy is a core-Sérsic galaxy. Moreover, there is a tendency for the AGN of core-Sérsic galaxies to be more radio-loud than those of the coreless galaxies (e.g. Capetti & Balmaverde 2005, 2006; Kormendy et al. 2009; Richings et al. 2011). Indeed, NGC 5322 is classified as a radio-loud galaxy (e.g. Baldi & Capetti 2009).

4.3.2 Kinematically decoupled core

The gas-poor major merger scenario of the spheroid (Section 4.3.1) cannot naturally account for the galaxy’s inner stellar disc which counter-rotates rapidly with respect to the slowly rotating spheroid (Bender 1988, see Section 3.4). Galaxies with central regions that are kinematically decoupled from the rest of the galaxy are thought to be common, but their formation mechanisms are still unclear (e.g. Efstathiou, Ellis & Carter 1982; Bender 1988; Franx & Illingworth 1988; Balcells & Quinn 1990; Hernquist & Barnes 1991; Forbes et al. 1995; Carollo et al. 1997; McDermid et al. 2006; Tsatsi et al. 2015). Integrating our Sérsic fit and using ε ∼ 0.4, we find an I-band M_I ∼ − 19.29 mag for the counter-rotating disc, dominated at all radii by the spheroid (Fig. 4). This disc only makes up ≈ 2.6 per cent of the total galaxy light and using a mass-to-light ratio M/L_I ∼ 2.8, the same as the spheroid’s M/L_I, gives a stellar disc mass of M_{*, disc} ∼ (6.4 ± 1.8)× 10⁹  M_⊙, somewhat smaller than the stellar disc mass from Bender (1990, |$M_{*,\rm disc} \sim 10^{10}\,\mathrm{M}_{\odot }$|⁠). The residual structure along the major axis of the galaxy due to the counter-rotating stellar disc (Figs 1 and 3) is in excellent agreement with the galaxy velocity map by Krajnović et al. (2011, their fig. C4).

Kinematically decoupled core (KDC) formation due to survived cores of accreted gas-poor satellites has been discussed by Kormendy (1984) and Balcells & Quinn (1990). These mechanisms are, however, inconsistent with this galaxy’s inner dust disc and central radio source which are a telltale signature of nuclear cold gas accretion events. It is unlikely that the KDC of NGC 5322 (i.e. the inner stellar disc) was built through major mergers of gas-rich or partially gaseous galaxies (e.g. Barnes & Hernquist 1991; Hernquist & Barnes 1991; Bender & Surma 1992; McDermid et al. 2006; Bois et al. 2011; Krajnović et al. 2011; Naab et al. 2014; Toloba et al. 2014; Krajnović et al. 2015; Tsatsi et al. 2015). These processes drive nuclear inflows of gas and subsequently trigger intense central starbursts, which are typically |${\sim } 1\hbox{ per cent}-5\hbox{ per cent}$| of the host galaxy light (Hopkins et al. 2008, 2009, see also Park, Smith & Yi 2017), resulting in a coreless galaxy (e.g. Faber et al. 1997). Furthermore, using numerical simulations, Hau & Thomson (1994) proposed a formation model for KDC galaxies, including NGC 5322, where a retrograde flyby interaction of a galaxy with another elliptical galaxy leads to a slowly rotating halo and a rapidly rotating KDC core. Their simulations assume an inner core plus an other halo components for the galaxies, but NGC 5322 has an inner stellar disc, a spheroid, and an outer halo (see Fig. 4 and Section 3).

The properties of NGC 5322 favour the formation of the KDC due to the accretion of a gas-rich satellite which settles at the galaxy centre.

4.3.3 A three-phase assembly for NGC 5322

We argue that a more likely scenario for NGC 5322 is a three-phase assembly, in which, the galaxy had its spheroid built earlier through (⁠|$\mathcal {N} \sim 2-7$|⁠) violent dry major mergers involving SMBHs (e.g. Faber et al. 1997; Kormendy 1999; Rest et al. 2001; Ravindranath et al. 2001; Trujillo et al. 2004; Lauer et al. 2005, 2007; Bell et al. 2006; Ferrarese et al. 2006; Man et al. 2012; Xu et al. 2012; Man, Zirm & Toft 2016), followed by the subsequent accretion of (a satellite with) metal-rich gaseous material which settles at the galaxy centre into a rapidly rotating (discy) cold gas reservoir. This, in turn, later fuels the central radio source and converts into the inner counter-rotating stellar disc (see Bertola, Buson & Zeilinger 1988; Franx & Illingworth 1988; Faber et al. 1997; Caon, Macchetto & Pastoriza 2000; Cappellari et al. 2002). Rix & White (1992) argued that this stellar disc is embedded in the dynamically hot spheroid (i.e. V_rot/σ ∼ 1.4). The outer halo (≈24.0 per cent of the total galaxy light) could naturally have grown later via dry minor mergers and the accumulation of stars stripped during flybys (e.g. Zolotov et al. 2009; Hilz et al. 2012; Hilz, Naab & Ostriker 2013; Hirschmann et al. 2015; Rodriguez-Gomez et al. 2016, their fig. 4). Due to NGC 5322’s group environment and therefore the lack of intracluster light (Garcia 1993; Madore et al. 2004), this galaxy’s halo and spheroid contribute equally in the outermost regions, compared to BCG haloes which typically dominate their host spheroids’ light at large radii (e.g. Seigar, Graham & Jerjen 2007; Donzelli, Muriel & Madrid 2011; Dullo et al. 2017).

4.3.4 Low-luminosity AGN feedback

A key challenge to our three-phase formation picture is that the new star formation due to the gas/gas-rich satellite accretion can refill the depleted core (see Faber et al. 1997). However, the core can be protected if the non-thermal AGN feedback swiftly quenches the in situ star formation after the stellar disc is formed. Serra et al. (2012) detected neutral hydrogen absorption in NGC 5322, although Hibbard & Sansom (2003) did not detect neutral hydrogen associated with the galaxy. Using e-MERLIN observations, we tentatively detected neutral hydrogen absorption near the central region of NGC 5322. Also, Spolaor et al. (2010) found a significant α enhancement at the central regions of the galaxy ([α/Fe] ∼ 0.30), suggesting short star formation time-scales (e.g. Matteucci & Tornambe 1987; Thomas, Greggio & Bender 1999; Thomas et al. 2005).

Our radio data analysis suggests that NGC 5322 houses a low-luminosity AGN (LLAGN). As noted in Section 3.5, the 1.5 GHz radio continuum emission of the galaxy has a core radio power and brightness temperature of P_core ∼ 6.55 × 10²⁰ W Hz⁻¹ and T_{B, core} ∼ 4.5 × 10⁷ K, respectively (Table 2), which agree with those those reported for low-luminosity AGN (T_{B, LLAGN} ≳ 10⁷ K, e.g. Nagar et al. 2002; Nagar et al. 2005; Nyland et al. 2013 and P_{core, LLAGN} ∼ 10¹⁸–10²⁵ W Hz⁻¹, e.g. Ho & Ulvestad 2001). For NGC 5322’s radio jets, we measured a low radio power of P_jets ∼ 7.04 × 10²⁰ W Hz⁻¹. Using Chandra X-ray spectra and the surface brightness profile of NGC 5322, Balmaverde, Baldi & Capetti (2008) argued that a low accretion rate (≈0.75 M_⊙ yr⁻¹) of hot gas on to the AGN at the bondi radius of 15 pc is enough to account for these radio jet’s kinetic power. Fig. 3 reveals that the residual image obtained by subtracting the ellipse model image of the galaxy exhibits two bright cones, tracing the orientations of the radio jets. However, since the jets have a low radio power, these bright cones are less likely to form due to radio mode AGN photoionization of the interstellar medium of the galaxy (e.g. Croton et al. 2006; Heinz et al. 2006; McNamara et al. 2009).

Furthermore, the modelling of the HST brightness profile shows that NGC 5322 has a normal size core for the galaxy’s σ and M_V, thus it is unlikely that a pre-existing large core in this galaxy was partially replenished by new stars (see Section 4.2).

4.3.5 Three-phase scenario versus galaxy properties and environment

The three-phase formation scenario reconciles well with the colour, stellar population properties, globular cluster system and the environment of NGC 5322. Spolaor et al. (2010, their tables C1–C3) reported a positive age gradient for the galaxy where the age rises from a central value of 7.4 Gyr to an average age of 12.6 Gyr. The central intermediate age (7.4 Gyr) stellar population (see also Proctor & Sansom 2002) is consistent with a stellar population that is a mixture of relatively young and old stars associated mostly to the stellar disc (i.e. KDC) and spheroid, respectively. NGC 5322’s negative radial metallicity gradient of ∼−0.20 dex decade⁻¹ is such that the galaxy is metal-rich in the central regions ([Z/H] ∼ 0.45] and it becomes less metal-rich with increasing radius (Spolaor et al. 2010). This negative metallicity gradient is expected since the small (gas-poor) satellites forming the outer halo are metal-poor. The higher central metallicity may reflect an early dissipative merger-induced starburst in the (coreless) progenitors of the spheroid and may typically remain intact during dry mergers (e.g. De Lucia et al. 2006; Spolaor et al. 2010) and the accretion of (a satellite with) already enriched gaseous material that built up the stellar disc. The galaxy becomes gradually bluer towards larger radii (see Section 3.1), as also reflected in the decline of the fraction of red globular clusters going from the centre to larger radii (Harris et al. 2006, their table 5 and fig. 24). In line with the three-phase picture, NGC 5322 resides in a poor galaxy group environment, which favours mergers (e.g. Zabludoff & Mulchaey 1998; Carlberg et al. 2001).

The nuclear dust disc and the inner stellar disc of NGC 5322 are both oriented along the major-axis of the galaxy, suggesting a common (external) origin for both these structures. The dust disc is nearly perpendicular to the radio jets (Section 3.5). However, Verdoes Kleijn & de Zeeuw (2005) did not find a trend between the radio jets and dust orientations for their sample of galaxies with nuclear dust disc, in contrast they found that radio jets tend to be perpendicular to the dust structure in the ‘dust lane’ galaxies. A number of past studies found a strong link between the presence of dust in galaxies and radio activities (e.g. Jaffe et al. 1994; van Dokkum & Franx 1995; Martel et al. 1999; Verdoes Kleijn et al. 1999; Tran et al. 2001; Verdoes Kleijn & de Zeeuw 2005; Lauer et al. 2005; Baldi & Capetti 2008; Shabala et al. 2012). The dust disc in NGC 5322 may signal the accretion of externally acquired gaseous material on to the the central supermassive black hole which feeds the AGN and powers the radio jets, extending ∼1.6 kpc at a PA of 172°.

5 CONCLUSIONS

As part of the LeMMINGs survey, we investigated the nuclear activity and central structure of the radio galaxy NGC 5322 using a new sub-arcsec L-band (1.25–1.75 GHz) e-MERLIN radio observation together with HST WFC3/IR, ACS/WFC and SDSS imaging. We extracted a composite (HST WFC3/IR and ACS/WFC + SDSS) major-axis surface brightness profile and isophotal parameters. The inner |$(R\lesssim 3\, \text{arcsec})$| part of this composite light profile was extracted from the redder HST WFC3/IR image, applying a pixel mask which mapped the galaxy’s nuclear dust disc with a radius of 1|${^{\prime\prime}_{.}}$|7 ≈ 246.5 pc and a thickness of 0|${^{\prime\prime}_{.}}$|6 ≈ 87 pc near the galaxy centre in the F814W/ACS/WFC image. For the first time we have decomposed NGC 5322’s composite light profile, covering R ∼ 200 arcsec, into an inner Sérsic (n ∼ 1 ± 0.2) stellar disc, a boxy core-Sérsic spheroid plus an outer, exponential stellar halo. Our primary findings are follows:

1. We have derived a central stellar mass deficit (M_def ∼ 1.1 × 10⁹ M_⊙) in the core-Sérsic spheroid with a core size R_b ∼ 0|${^{\prime\prime}_{.}}$|37 ± 0|${^{\prime\prime}_{.}}$|04 ≈ 53.7 ± 5.4 pc and n ∼ 4 ± 0.8. This core size is in good agreement with those estimated utilizing the spheroid’s σ ∼ 229 km s⁻¹ and M_V ∼ −21.77 mag ( ≈ 73.4 per cent of the total galaxy light) along with the R_b–σ and R_b–M_V relations for core-Sérsic galaxies (Dullo & Graham 2014), i.e. |$R_{\rm b, \sigma \text{-}based} \approx 44.0 \pm 6.1$| pc and |$R_{{\rm b},M_{V}\text{-}{\rm based}} \approx 48.6 \pm 7.2$| pc. We also measured low F110W-band extinction (A_F110W ∼ 0.031–0.047 mag) in the regions immediately surrounding the masked regions near the galaxy centre, compared to our adopted uncertainty on the surface brightness μ_b, μ_{e, S}, and μ_{0, h}, i.e. ∼ 0.02 mag arcsec⁻². The good agreement in R_b and the low IR extinction give us confidence that our measurement of NGC 5322’s R_b (and therefore M_def) is not plagued by the effects of the nuclear dust disc. We found a stellar mass M_* ∼ (6.4 ± 1.8) × 10⁹ M_⊙for the stellar disc, somewhat smaller than that reported in the past.

2. We find that NGC 5322 has a core 1.5 GHz radio continuum emission with a power of P_core ∼ 6.55 × 10²⁰ W Hz⁻¹ and a brightness temperature of T_{B, core} ∼ 4.5 × 10⁷ K (Table 2). Both these figures favour a non-thermal nuclear radio emission due to a low-luminosity AGN. For the galaxy’s radio jets, we measured a low radio power of P_jets ∼ 7.04 × 10²⁰ W Hz⁻¹.

3. We propose a three-phase formation scenario to explain the formation and evolution of NGC 5322. The spheroid may have built up early through (2–7) gas-poor major mergers, as traced by the M_def/M_BH ratio. The galaxy subsequently cannibalized a satellite containing metal-rich gas, with stars and gas settling at the centre and forming the rapidly rotating stellar disc, counter-rotating with respect to the spheroid. The outer stellar halo later grew inside-out via minor mergers and accretion of tidal debris. Our formation model for NGC 5322 is consistent with the galaxy’s colour, stellar population properties, globular cluster system, and environment.

4. NGC 5322’s dust disc oriented along the galaxy’s major-axis is nearly perpendicular to the radio jets, which extend ∼1.6 kpc at a PA of 172°. Nuclear dust features may indicate the accretion of externally acquired gaseous material on to the AGN, powering central radio sources. For NGC 5322, this radio mode (low-luminosity) AGN feedback that drives the low power jets might have quenched the late-time star formation in the galaxy after the stellar disc is formed, preventing the refill of the depleted core.

Acknowledgements

We thank the referee for their careful reading of the paper and many constructive suggestions that improved the paper. BTD acknowledges support from a Spanish postdoctoral fellowship ‘Ayudas para la atracción del talento investigador. Modalidad 2: jóvenes investigadores, financiadas por la Comunidad de Madrid’ under grant number 2016-T2/TIC-2039. BTD acknowledges financial support from the Spanish Ministry of Economy and Competitiveness (MINECO) under grant number AYA2016-75808-R. BTD and JHK acknowledge financial support from MINECO under grant number AYA2013-41243-P. JHK acknowledges financial support from the European Union’s Horizon 2020 research and innovation programme under Marie Skłodowska-Curie grant agreement No 721463 to the SUNDIAL ITN network, and from the Spanish Ministry of Economy and Competitiveness (MINECO) under grant number AYA2016-76219-P. IMcH thanks the Royal Society for the award of a Royal Society Leverhulme Trust Senior Research Fellowship. RDB and IMcH also acknowledge the support of STFC under grant [ST/M001326/1]. Based on observations made with the NASA/ESA Hubble Space Telescope, and obtained from the Hubble Legacy Archive, which is a collaboration between the Space Telescope Science Institute (STScI/NASA), the Space Telescope European Coordinating Facility (ST-ECF/ESA), and the Canadian Astronomy Data Centre (CADC/NRC/CSA). e-MERLIN is a National Facility operated by the University of Manchester at Jodrell Bank Observatory on behalf of STFC. Funding for the Sloan Digital Sky Survey IV has been provided by the Alfred P. Sloan Foundation, the U.S. Department of Energy Office of Science, and the Participating Institutions. SDSS-IV acknowledges support and resources from the Center for High-Performance Computing at the University of Utah. The SDSS web site is www.sdss.org.

SDSS-IV is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS Collaboration including the Brazilian Participation Group, the Carnegie Institution for Science, Carnegie Mellon University, the Chilean Participation Group, the French Participation Group, Harvard-Smithsonian Center for Astrophysics, Instituto de Astrofísica de Canarias, The Johns Hopkins University, Kavli Institute for the Physics and Mathematics of the Universe (IPMU) / University of Tokyo, Lawrence Berkeley National Laboratory, Leibniz Institut für Astrophysik Potsdam (AIP), Max-Planck-Institut für Astronomie (MPIA Heidelberg), Max-Planck-Institut für Astrophysik (MPA Garching), Max-Planck-Institut für Extraterrestrische Physik (MPE), National Astronomical Observatories of China, New Mexico State University, New York University, University of Notre Dame, Observatário Nacional / MCTI, The Ohio State University, Pennsylvania State University, Shanghai Astronomical Observatory, United Kingdom Participation Group, Universidad Nacional Autónoma de México, University of Arizona, University of Colorado Boulder, University of Oxford, University of Portsmouth, University of Utah, University of Virginia, University of Washington, University of Wisconsin, Vanderbilt University, and Yale University.

Footnotes

1

http://leda.univ- lyon1.fr

2

https://hla.stsci.edu/hlaview.html

3

https://archive.stsci.edu

4

http://www.sdss.org/dr13/

5

Dullo & Graham (2012, 2014) discussed the core-Sérsic model in detail and provided a comparison between the model and the Nuker model (Lauer et al. 1995), see also (Graham et al. 2003; Trujillo et al. 2004; Ferrarese et al. 2006).

6

http://www.aips.nrao.edu/

7

We use an I-band absolute magnitude for the Sun of M_{⊙, I} ∼ 4.14 mag (http://www.ucolick.org/cnaw/sun.html).

8

A caveat here is that the ‘core’ (Lauer et al. 1995; Faber et al. 1997) and ‘core-Sérsic’ (Graham et al. 2003) structural classifications are not identical. Unlike ‘core-Sérsic’ galaxies, ‘core’ galaxies identified by the Nuker model (e.g. Lauer et al. 1995) do not necessarily contain partially depleted cores relative to their spheroids’ outer Sérsic profile (Dullo & Graham 2012, 2013, 2014, 2015).