An extreme case of galaxy and cluster co-evolution at z = 0.7

A search of public radio surveys (TGSS, GLEAM, WISH, NVSS, and VLASS) at the location of the BCG of eMACS J0252 reveals a relatively bright, compact radio source detected from 76 MHz to 3,GHz.

A short SCUBA2 observation of eMACS J0252 was obtained at 850 $\mu$m in August 2015 as part of a larger survey of eMACS clusters (PI Edge). The exposure of 50 min in good conditions ($\tau_{225}=0.06$) achieved an rms sensitivity of 4.2 mJy per beam; the lack of a detection in this observation places 3-$\sigma$ upper limits of 12.6 mJy on the BCG flux density at 850 $\mu$m.

The spectroscopic observations described in Section 2.3 yield a heliocentric cluster redshift of $z=0.7028$ and a velocity dispersion of $\sigma=1020^{+180}_{-190}$ km s^-1, based on the 49 concordant redshifts shown in Fig. 2 and listed in the Appendix.

We find no significant substructure along the line of sight in the overall redshift distribution (Fig. 2), as would be expected if the observed velocity dispersion were inflated by ongoing or recent merger activity perpendicular to the plane of the sky.

The large clustercentric distances reached in our DEIMOS observations enable us to probe the large-scale environment of eMACS J0252 (Fig. 2) and test whether the system is isolated in three dimensions or connected to mass concentrations in the vicinity. However, the redshift distribution of the galaxies observed to date exhibits no clear trends with either right ascension or declination.

At the cluster redshift of $z=0.703$, the X-ray flux detected in the RASS corresponds to an X-ray luminosity of $(1.2\pm 0.4)\times 10^{45}$ erg s^-1 (0.1–2.4 keV). Although, at face value, this measurement suggests that eMACS J0252 is one of a mere handful of exceptionally X-ray luminous clusters known at $z>0.6$, this assessment remains subject to significant uncertainty, owing to the RASS detection in the direction of the source consisting of a mere 12 net photons. A robust determination of the X-ray properties of eMACS J0252 will have to account for contamination from X-ray point sources, a task that requires much better angular resolution than is achieved in past or current X-ray all-sky surveys (ROSAT and eROSITA).

Scrutiny of the spectra obtained with MUSE (Section 2.3) reveals unambiguous signs of strong gravitational lensing in the form of two multiple-image systems. The giant arc almost due West of the BCG is found to consist of two merged images of a galaxy at $z=3.458$. We show the MUSE spectrum of the arc in Fig. 3 (left); features anchoring the redshift measurement are faint Ly$\alpha$ emission spatially offset from the continuum (and included as subsystem 1a in our modeling of the mass distribution), as well as Si ii and C ii absorption. A second multiple-image system is spectroscopically identified through pronounced Ly$\alpha$ emission at 5420Å, revealing the source redshift to be $z=4.148$ (Fig. 3, right). The components of all multiple-image systems are listed in Table 1 and shown and labeled in the insets to Fig. 1; for Systems 1a and 2, which are barely detected in the HST SNAPshot of eMACS J0252, we mark the image positions by overlaying contours of the Ly$\alpha$ flux from our MUSE observations.

We use Lenstool (Jullo et al., 2007) to create a lens model of eMACS J0252. Adopting double Pseudo Isothermal Elliptical (dPIE) profiles for the mass distributions of both cluster- and galaxy-scale components, we optimise the shape and mass scaling parameters of the cluster-scale component and the central BCG to reproduce the three multiple-image systems. We include all spectroscopically confirmed cluster members as galaxy-scale halos that act as local perturbers of the large-scale gravitational potential. The best-fit model, characterized by the parameters listed in Table 2, yields an rms offset of 0.09″ between the predicted and observed locations of all images. Like the distribution of optical light from the BCG, the mass distribution of eMACS J0252 is characterized by an exceptionally high ellipticity; our lens model requires an axis ratio of $b/a=0.34\pm 0.01$ or an ellipticity of $e=0.79$,¹¹1We define ellipticity as $e=\frac{1-(b/a)^{2}}{1+(b/a)^{2}}$, with $b/a$ being the axis ratio. a value far more extreme of the average of 0.52 found, e.g., from statistical gravitational-lensing analyses of SDSS clusters (Shin et al., 2018).

The critical lines (for a source at $z=3.458$) from the best-fitting lens model are shown in Fig. 4. The projected mass within a radius of 250 kpc is $(1.94\pm 0.02\pm 0.1)\times 10^{14}$ M_⊙ (statistical and estimated systematic uncertainties are listed separately); the total mass (not well constrained by strong-lensing measurements) exceeds $5\times 10^{14}$ M_⊙. With an effective Einstein radius²²2We define the effective Einstein radius as $\sqrt{A_{\rm crit}/\pi}$, where $A_{\rm crit}$ is the area enclosed by the critical line. of 19.9″ for System 1 at $z=3.458$ (15.5″ for a source at $z=2$), eMACS J0252 is a moderately strong gravitational lens, in agreement with the general trend for more powerful cluster lenses to be disturbed active mergers, extreme cases being MACSJ0717.5+3745 (e.g., Limousin et al., 2016) or Abell 2744 (e.g., Jauzac et al., 2015).

The cD galaxy ($\alpha_{\rm J2000}=$ 02^h 52^m 27.25^s, $\delta_{\rm J2000}=-$21^∘ 00^′ 51.2″; Figure 1) in the core of eMACS J0252 is characterized by a prodigious luminosity ($M_{\rm g}=-24.1$), intersecting dust lanes, and an overall morphology that shows a strong resemblance to BCGs in cool-core clusters at lower redshift (Donahue et al., 2015). The archival photometry from PanSTARRS-1, GALEX and WISE yields a luminosity of $5\times 10^{11}$ L_⊙ for the BCG in the rest-frame $i$-band and near-UV, optically blue colours similar to those of line-emitting BCGs at lower redshift (Green et al., 2016).

Even more remarkable is the BCG’s extreme elongation (see Figure 1 and 4). Its ellipticity of $e=0.78$ (derived from an axis ratio of 0.35 measured from HST imaging) is unrivaled among similarly massive clusters (Herbonnet et al., 2019) and a $>3\sigma$ outlier compared to the average axis ratio of $0.7\pm 0.1$ established for BCGs in general (Donzelli et al., 2011). Although the light envelopes of BCGs are known to become more elliptical with increasing radius, axis ratios below 0.4 are extremely rare and only observed at radii of at least 100 kpc (Kluge et al., 2020), far larger than the approximately 40 kpc probed by our shallow HST data (Fig. 6).

The most notable feature of the bright compact radio source associated with the BCG is a very steep lower-frequency component ($\alpha\sim-2.0$) that dominates below 300 MHz, and whose spectral slope is comparable to that of the massive lobed source in MS 0735+74 ($z=0.21$; Cohen et al., 2005); it also shares with MS 0735+75 a similar rest-frame radio power of $1.3\times 10^{25}$ W Hz^-1. The radio emission above 1 GHz has a significantly flatter spectrum ($\alpha\sim-0.5$), comparable to that of the compact-core emission found in all line-emitting BCGs (Hogan et al., 2015) (but notably missing in MS 0735+74). In terms of radio power ($7.3\times 10^{24}$ W Hz^-1 at 10 GHz), the core emission of the BCG in eMACS J0252 falls into the top 10 percentile of the distribution compiled by Hogan et al. (2015) while the more extended, steep-spectrum emission still ranks in the top third. These properties suggest that accretion onto the active nucleus on the BCG is ongoing at present, after a sustained period of activity in the past that powered the aged, steep-spectrum lobes. A low-frequency observation at higher resolution of this source would help constrain the timing and energetics of this past outburst.

Strong absorption features, such as Ca ii (H+K), the G band, and Mg i b, in the spectrum of the BCG, as well as a pronounced Balmer break in the continuum, unequivocally establish the presence of an old stellar population, as found in all elliptical galaxies (Fig. 5). Rare in average ellipticals (but not uncommon in BCGs; Green et al., 2016, and references therein) is the pronounced line emission seen in Fig. 5.

From the [O ii]$\lambda 3727$Å emission-line flux recorded in our MUSE observation we measure a high star-formation rate of 85 M_⊙ yr^-1, following Kennicutt (1989). Although this value accounts for both Galactic extinction and intrinsic extinction as estimated from the observed Balmer decrement, the latter is flux-weighted and thus biased against high-extinction regions. Indeed the corresponding $E(B-V)$ value of about 0.05 appears unphysically low for the derived high intrinsic [O ii]$\lambda 3727$Å emission-line luminosity of $6.1\times 10^{42}$ erg s^-1, a combination that is in stark conflict with the correlation found by Kewley et al. (2004). Considering, in addition, the compelling evidence of a dust lane (apparent in the short HST SNAPshot shown in Fig. 1), we therefore consider the quoted intrinsic extinction and the corresponding star-formation rate lower limits. The lack of a detection in the SCUBA2 observations, however, places a corresponding 3-$\sigma$ upper limit of 300 M_⊙ yr^-1 on the BCG star-formation rate. Both limits are shown in Fig. 8 and identify the BCG of eMACS J0252 as an extreme system in terms of star-formation rate.

We note that a contribution to the observed [O ii] flux from nuclear activity can not be ruled out. Not only do the system’s radio properties suggest the presence of an AGN, the ratio of the fluxes of the [O iii] and H$\beta$ emission lines of $\log([\text{O\,{iii}}]/{\rm H}\beta)=-0.25$ also places the BCG of eMACS J0252 in the composite region of the BPT diagram (Baldwin et al., 1981). A correction for any AGN contribution to the [O ii] flux (and hence the deduced star-formation rate) will require near-infrared spectroscopy to determine the H$\alpha$ emission-line flux.

Using the pPXF code of Cappellari & Emsellem (2004), we measure the central stellar velocity dispersion of the BCG to be ($380\pm 30$) km s^-1, well above the average value of 260 km/s for BCGs in the local Universe (Lauer et al., 2014) and high for BCGs at any redshift (Sohn et al., 2020).

The strong [O ii] emission from the BCG (Fig. 5) extends to over 20 kpc in radius from the galactic centre, as shown in Fig. 6. In order to constrain the BCG’s kinematics, we track the spectral location and shape of the [O ii] line across the full extent of this region. We interpret the recorded shifts of the centroid of the [O ii] line as being caused by changes in the radial component of the peculiar velocity of the ISM, noting that variations in the ratio of the strengths of the 3726.0Å and 3728.7Å lines that make up the [O ii] doublet³³3Such variations would be indicative of changes in the electron density (Osterbrock, 1974). can not account for the observed spectral shifts. The resulting map of the BCG’s radial-velocity field is shown in Fig. 7. Unlike some line-emitting BCGs at lower redshift (Hamer et al., 2016), the BCG of eMACS J0252 shows no pronounced velocity asymmetry indicative of rotation or bulk motions associated with “sloshing" (first described by Markevitch et al., 2000), but instead a radial gradient of $(90\pm 30)$ km s^-1 amplitude that is consistent with radial infall of gas: the highest (positive) radial velocities are observed near the galactic centre, and no motion is detected at the largest radii (in projection) where infall is perpendicular to our line of sight. The timescale for infall from 30 kpc at this velocity is 330 Myr, in good qualitative agreement with the cooling times of about 1 Gyr observed for fully relaxed cool-core clusters (e.g., Hudson et al., 2010).

The above interpretation of the radial-velocity pattern shown in Fig. 7 as being indicative of isotropic infall requires the gas within the shown region to be optically thick⁴⁴4If the gas around the BCG were optically thin, like the ICM over most of the cluster volume, [O ii] emission would be detectable both from gas in front and behind the BCG; as a result, isotropic infall would create no net shift of the [O ii] line but only an effective broadening.. Although speculative for eMACS J0252 at this point, this assumption is consistent with the high reddening values of $E(B-V)\approx 0.6$ measured in the centres of massive cool-core clusters (O’Dea et al., 2010; McDonald et al., 2013) and excess absorption equivalent to neutral-hydrogen column densities of several $10^{21}$ cm^-2 (e.g., McDonald et al., 2019).