EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 676 (August 2023) A&A, 676 (2023) A39 Full HTML
Open Access
Issue
A&A
Volume 676, August 2023
Article Number A39
Number of page(s) 14
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/202346580
Published online 04 August 2023

© The Authors 2023

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1. Introduction

According to the standard cosmological paradigm Λ-cold dark matter (ΛCDM; Peebles & Ratra 2003), clusters of galaxies are the latest and most massive structures to form in the Universe. Their process of formation and evolution follows a hierarchical structure through the accretion of matter in filaments of the cosmic web, the addition of nearby galaxies, and the merging with groups and other clusters. Thus, their cluster formation pathway is plagued with interactions between galaxies and also with the intracluster medium. As a direct consequence of these violent processes, part of the stars in the interacting galaxies is freed to the intracluster space and is just bound by the gravitational potential of the cluster. This composes the so-called intracluster light (ICL). The ICL appears as an extended diffuse light in between the galaxy members of the clusters and is especially concentrated around the brightest cluster galaxy (BCG), with a typical low surface brightness of μV > 26.5 mag arcsec−2 (e.g., Rudick et al. 2006). Thus, the ICL properties, from morphology to stellar composition, are intimately linked to the characteristics of its progenitor galaxies and can provide us with valuable information about the dynamics that govern the system as a whole. For example, it is estimated that during major galaxy mergers with the BCG, which generally occur at z > 1, approximately between 30% and 50% of the stars of the merging body become unbound and end up in the ICL (Murante et al. 2007; Conroy et al. 2007; Lidman et al. 2012; Joo & Jee 2023). As a consequence, the color and metalicity radial profiles of the ICL are flat (or have a very shallow gradient) in color and metallicity, reflecting the mixture of stellar populations that have been thrown into the intracluster space. On the other hand, at intermediate and low redshifts (z < 0.5), the main mechanisms of ICL formation are thought to be cluster-cluster mergers, tidal stripping of luminous galaxies, shredding of dwarf galaxies, and preprocessing in infalling groups (e.g., DeMaio et al. 2015; Melnick et al. 2012; Rudick et al. 2006, 2011; Puchwein et al. 2010; Murante et al. 2007; Conroy et al. 2007; Montes & Trujillo 2014, 2018; Morishita et al. 2017; Jiménez-Teja et al. 2018, 2019, 2021; de Oliveira et al. 2022; Ragusa et al. 2023). These processes cause color gradients in the ICL: (1) Massive galaxies first loose the bluer stars located in their outer layers, and as they orbit toward the potential well of the cluster or continue to interact with another galaxies, they progressively free the redder stars of their inner regions. (2) Patches of ICL appear in regions in which groups enter the gravitational potential of the cluster (e.g., Iodice et al. 2017). Finally, (3) gravitational forces can act on dwarf galaxies at different clustercentric radii and cause a bluer ICL in the outskirts of the cluster because lower-mass dwarfs experience the gravitational effect of the potential of the cluster at larger radii than more massive and more metal-rich dwarfs (DeMaio et al. 2018). Observational evidence has indicated that the projected number density of dwarf galaxies is anticorrelated with clustercentric distance, which may be the consequence of this transfer of material to the ICL (e.g., Venhola et al. 2018). Moreover, observations of nearby clusters confirmed that red low surface brightness dwarfs are more concentrated in the center of clusters, show more indications of disturbance, and tend to have lower apparent axis ratios than normal dwarfs. This is also consistent with the cluster tidal forces acting on them and relocating their stars to the ICL (Lim et al. 2020; Venhola et al. 2022).

Although different mechanisms have been identified as primary sources at different redshifts, only a few works in the literature have indeed analyzed the ICL of clusters at high redshift. It is thought that the ICL formation shows a burst at z < 0.5 (Montes & Trujillo 2018), but abundant ICL is already set up at z > 1 (e.g., Burke et al. 2012; Ko & Jee 2018; Joo & Jee 2023) and even at z > 2 (Adami et al. 2013). Different mechanisms are claimed to be primary sources depending on the redshift. The ICL formation is mostly linked to BCG build-up at z > 1 and to other channels related to satellite galaxies and infalling groups or galaxies at z < 1. Interestingly, this two-phase formation scenario implies that the ICL properties strongly depend on the instantaneous dynamical state of the cluster (Jiménez-Teja et al. 2018). This opened a new window for inferring the merging stage of clusters and even the age of particular systems, which are called fossil groups (Dupke et al. 2022), as long as multiwavelength observations are available.

Here, we perform a multiwavelength study in the optical and infrared of the ICL fraction and in X-rays of the intracluster gas of a high-redshift cluster observed by the Reionization Lensing Cluster Survey (RELICS; Coe et al. 2019), SPT-CLJ0615−5746 at z = 0.972 (SPT0615 hereafter). This cluster has been one of the very few high-redshift clusters that were controversially classified as relaxed by some previous works (Planck Collaboration XXVI 2011; Bartalucci et al. 2019; Connor et al. 2019).

This paper is organized as follows. We first describe the optical and infrared data available for this clusters, the data reduction process, and the algorithms we used to derive the ICL maps and fractions in Sect. 2. Next, we carry out an extensive analysis of the available X-ray data of SPT0615 in Sect. 3. We discuss the results of the combined optical, infrared, and X-ray analysis in Sect. 4 and draw our main conclusions in Sect. 5. Throughout this paper, we assume a standard ΛCDM cosmology with H0 = 70 km s−1 Mpc−1, Ωm = 0.3, and ΩΛ = 0.7. All magnitudes refer to the AB system.

2. Optical and infrared data and analysis

The cluster SPT0615 (RA = 6h15m56 . s $ {{\overset{\text{s}}{.}}} $27, Dec = −57°45′50″ [J2000.0]) can be also found in the literature as PSZ1 G266.56−27.31. It is one of the most distant and massive clusters detected by Planck at z = 0.972 with M 500 ∼ 6 . 77 − 0.54 + 0.49 × 10 14 M ⊙ $ M_{500}\sim 6.77^{+0.49}_{-0.54}\times 10^{14}\,M_{\odot} $, as estimated by this collaboration (Planck Collaboration XXVII 2016). SPT0615 was independently discovered by both the South Pole Telescope survey (Williamson et al. 2011) and the Planck Collaboration (Planck Collaboration XXVI 2011) using the Sunyaev–Zel’dovich (SZ) effect. It is exceptionally luminous, with a [0.1−2.4] keV band luminosity of (22.7 ± 0.8)×1044 erg s−1, and it is hot, with TX ∼ 11 keV (Planck Collaboration XXVI 2011).

We calculated ICL fractions (defined as the ratio of the ICL to the total light of the cluster in a certain filter) for SPT0615 in six optical and infrared bands. To do this, we built an ICL map and an image that only contained the members of the cluster and the ICL for each filter considered. We used data gathered by the Hubble Space Telescope (HST) and the algorithm called CHEFs (from Chebyshev-Fourier) intracluster light estimator (CICLE; Jiménez-Teja & Dupke 2016), which has been successfully applied before to high-quality HST data (Jiménez-Teja et al. 2018, 2021; de Oliveira et al. 2022; Dupke et al. 2022) to build the ICL maps, and a machine learning algorithm (Lopes & Ribeiro 2020) to determine the cluster membership.

2.1. HST data

RELICS (Coe et al. 2019) is a multi-orbit Hubble Treasure Program devoted to observing the 21 most massive and distant galaxy clusters, according to Planck estimates (Planck Collaboration XXVII 2016), as well as 20 additional systems selected by their strong-lensing nature. RELICS provides information in the optical with the Advanced Camera for Surveys (ACS) and in the infrared with the Wide Field Camera 3 (WFC3). Before the RELICS survey, archival HST images of SPT0615 existed in the optical filters F606W and F814W (programs 12477, PI: High; and 12757, PI: Mazzotta). RELICS completed these observations with one HST orbit in the F435W filter and two orbits split among the four infrared filters F105W, F125W, F140W, and F160W.

All images were reduced by the RELICS collaboration by applying the standard pipelines CALACS1 for the optical data and CALWF32 for the infrared images. These pipelines include corrections for bias, dark, flat-fielding, bias-striping, crosstalk, and charge transfer efficiency. Moreover, persistence models were calculated to mask the pixels in the infrared images in which some afterglow might remain from previous pointings. Although it is not optimal for low surface brightness studies, we showed in Jiménez-Teja et al. (2021) that the standard flat-field correction provided by the standard HST pipelines does not affect the resulting ICL fractions if total fluxes are measured. However, for images for which we observed that the standard MAST flat-field image might compromise the ICL measurement, we calculated new sky-flats using the prescriptions of Borlaff et al. (2019). Individual exposures where later aligned and combined using Astrodrizzle (Koekemoer 2002) to a pixel scale of 0.06 arcsec. We estimated the limiting surface brightness in 3 × 3 arcsec2 boxes for each filter using the prescription of Román et al. (2020). These values are shown in Table 1.

Table 1.

Limiting surface brightness calculated in boxes of 3 × 3 arcsec2, ICL fractions computed with CICLE, and breakdown of the ICL fraction error into the three sources considered.

2.2. Generation of the ICL maps

The code CICLE (Jiménez-Teja & Dupke 2016) eliminates the galactic luminous contribution by masking the stars and fitting the galaxies. Stars are masked either using a SExtractor segmentation map (Bertin & Arnouts 1996) or manually, when the pixels assigned by SExtractor to the star clearly do not cover their full extension. Galaxies are modeled using a combination of Chebyshev rational functions and Fourier series, which form a mathematical orthonormal basis capable of fitting objects with any type of morphology when they are smooth enough (Jiménez-Teja & Benítez 2012). For instance, saturated stars or objects located too close to the border images are excluded from the set of morphologies that CHEFs can fit. The CHEF algorithm determines the extension of the stellar halo of the galaxies modeled by estimating the point where the projected galactic surface either submerges into the sky or converges asymptotically to a certain value in each radial direction. The ICL tends to follows the gravitational potential of the cluster, so it is usually more concentrated around its center, close to the BCG. As a consequence, this procedure of finding the observational limits of the stellar halos is only valid for galaxies that are located relatively far from the cluster center. Thus, it cannot be applied to distinguish the BCG from the ICL because their projected distributions are usually aligned. For this particular case, CICLE used a curvature parameter, called minimum principal curvature (Patrikalakis & Maekawa 2010), to determine the limits of the region in which the BCG dominates the ICL. Intuitively, this parameter quantifies the change in the slope of a surface, point by point. CICLE calculates the curvature map of the composite BCG+ICL surface and identifies the maxima values (i.e., the points where it finds the highest change in the slope) with the transition from the BCG- to the ICL-dominated area. The more dissimilar the profiles or the steepness of the BCG and the ICL, the easier the detection of this transition. This does not mean, however, that the whole stellar halo of the BCG is contained within this region, but a small amount of it will still extend over the ICL-dominated region, covered by the ICL in projection. The BCG flux missed in this way is included in the final error budget as part of the so-called geometrical error, as described in Sect. 2.4. After modeling the BCG within this region, we obtain a map that only contains ICL and background.

Although the sky is already removed in calibrated images, it is mandatory to perform a refinement of this background in a final step. Many different approaches for this estimation of the background have been published, such as using nearby fields (observed under the same observational and technical characteristics; e.g., Jiménez-Teja & Dupke 2016), analyzing the distribution of the pixels that are considered as ICL-free (Morishita et al. 2017; Jiménez-Teja et al. 2021), or using specialized software such as SExtractor (e.g., Burke et al. 2012), or, more recently, NoiseChisel (Borlaff et al. 2019). To our understanding, the first approach would be the safest, but parallel fields with the same observational characteristics are not usually available. Studying the distribution of the blank pixels in the image is not feasible in our case because the background varies locally (even though we recalculated the sky flats, as decsribed in Sect. 2.1) and the field of view associated with some of the filters considered in this work is larger. Borlaff et al. (2019) made an exhaustive study of the different ways of estimating the sky and concluded that although all algorithms tend to overestimate the sky level, NoiseChisel (Akhlaghi & Ichikawa 2015) performed best. More recent works with different datasets have led to similar conclusions (Haigh et al. 2021; Kelvin et al. 2023). We used the updated version of NoiseChisel (Akhlaghi 2019) to calculate our background maps. Then, we removed them from the ICL+background images that were calculated previously to obtain the final ICL maps. In Fig. 1 we show the contours of the ICL in the six optical and infrared filters analyzed in this work, superimposed on the original images. The morphology of the ICL in the different bands is similar overall. It is elongated along a north-south axis with a position angle of about 30°. It also appears to be preferentially concentrated around two main clumps north and southwest of the BCG.

thumbnail Fig. 1.

ICL isocontours superimposed on the original images in the six HST ACS and WFC3 filters. For each band, we plot ten logarithmically spaced isocontours. The lowest level is calculated from the detection limit of the ICL (where it converges with the background level), and the highest level corresponds to its maximum value. We report the surface brightness limits for the lowest and highest isocontours inside each panel. Two main ICL clumps appear in all bands. They are more clearly separated in the bluer filters.

2.3. Machine-learning approach to photometric membership

The second step was to generate an image with the stellar content of the cluster, that is, the ICL and the cluster galaxies. We just added the CHEF models of the galaxies to the ICL map after determining the galaxies that belong to the cluster. We applied the code called reliable photometric membership (RPM; Lopes & Ribeiro 2020) to select galaxy cluster members based only on the photometric parameters of the galaxies lying along the line of sight of the clusters. RPM employs a machine-learning (ML) approach to derive membership probabilities, which are then used to derive a membership classification. Lopes & Ribeiro (2020) verified the code efficiency for low-redshift systems within R200. Jiménez-Teja et al. (2021) demonstrated that the code also works for higher-z systems (z < 1) within the RELICS survey footprint.

The ML method was trained and evaluated with galaxies within the regions of 18 CLASH clusters, covering a broad redshift range (0.0792 < z < 0.8950). The best results we achieved considered galaxies within 1.50 h−1 Mpc of the cluster center and with 15 ≤ F814W ≤ 25. Our final dataset (used to train and validate the ML method) comprises 927 galaxies with spectroscopic redshifts from the CLASH clusters. Further details of these steps can be found in Jiménez-Teja et al. (2021).

As discussed in Jiménez-Teja et al. (2021), we found that the stochastic gradient boosting (GBM) method results in a slightly better performance than other ML models. This is quantified by the completeness (also called true positive rate, TPR, or sensitivity) and purity (known as precision or positive predictive value, PPV). These two parameters track the relation of the sample of objects that are classified as members and the true population of members.

The photometric parameters employed by the ML model are Δ (F435W − F814W), Δ (F606W − F814W), Δ (F105W − F140W), Δ (F814W − F125W), Δ (F814W − F140W), LOG Σ5, and Δz phot. Δ stands for the offset relative to the mean magnitude, color, or cluster redshift. We did not use apparent magnitudes or observed colors because the redshift range of the training sample is broad. We also avoided absolute magnitudes and rest-frame colors that may be subject to large uncertainties associated with the photo-z precision and due to the k- and e-corrections.

We achieve high values of completeness (C) and purity P, C = 93.5%±2.4% and P = 85.7%±3.1%. This method was successfully applied to all galaxies in the regions of the 25 CLASH clusters, as well as for 35 of the 42 RELICS clusters, including WHLJ013719.8–082841 (for which the ICL was investigated in Jiménez-Teja et al. 2021). For the particular case of SPT0615, 174 galaxies are photometrically classified as cluster members. Additionally, we have spectroscopic redshifts for 58 galaxies (Connor et al. 2019) and 47 of them have a similar redshift as the cluster (velocity offsets smaller than 5000 km s−1). We then used the shifting gapper technique (Fadda et al. 1996; Lopes et al. 2009) to select members and exclude interlopers in the projected phase space. This is simply based on the application of the gap-technique in radial bins from the cluster center (described in Katgert et al. 1996), to identify gaps in the redshift (velocity) distribution. Instead of adopting a fixed gap, such as 1000 km s−1, we considered a variable gap, called density gap (Adami et al. 1998; Lopes et al. 2009), which depends on the number density of galaxies in the cluster region. Thus, according to the spectroscopic information, we found that 39 galaxies are members of SPT-CLJ0615–5746, while 8 galaxies are interlopers. Based on the photometric membership, 32 out of these 39 are also selected as members (∼82%). This provides a lower limit to our completeness because we have many more galaxies with photometry than spectra in this region.

One interesting point in the study of SPT0615 is the possible foreground structure at z ∼ 0.4, as pointed out by Paterno-Mahler et al. (2018). To investigate this, we explored the photometric redshift distribution of the galaxies in the region of SPT0615 (Fig. 2 top), and we verified that it is very likely that a second system is projected along the line of sight at zphot ∼ 0.44. Hence, we decided to apply the photometric membership selection and also take this foreground structure into account. We repeated the procedure described above for all the galaxies in the footprint of SPT0615, but now considering the centroid and photometric redshift of this possible foreground group as reference. We calculated two different probabilities for each galaxy to either belong to SPT0615 or to the structure at zphot ∼ 0.44. For each object, we included the highest membership probability of these two in our final catalog and classified it as member or interloper relative to the high-z cluster and to the foreground group. The color-magnitud diagram (CMD; see Fig. 2 bottom), shows a possible second red sequence with three brightest galaxies of very similar magnitude. Two of them are very close, southwest of SPT0615 BCG. We took the middle between these two galaxies as the center of the possible second structure. If it exists, 38 members would be identified in the foreground cluster or group. We inspect the spatial distribution of this putative foreground group in Sect. 4.

thumbnail Fig. 2.

Photometric cluster membership. Top: photometric redshift distribution of the galaxies in the region of SPT0615. Bottom: color-magnitude diagram with the members of SPT0615 at z = 0.97 in red and those of the possible foreground structure at z ∼ 0.44 in black. Gray points correspond to interlopers that do not belong to any of the two structures.

2.4. Calculation of the ICL fraction and error

We defined the limit of the ICL as the points in which the ICL is submerged into the sky level. We first calculated the radial profiles of the ICL and the background and searched for the average radius at which they coincide. For the sake of illustration, we plot these radial profiles (along with the profile of the total cluster) for the F160W band in Fig. 3. The errors are drawn from a jacknife resampling. The radial profile of the ICL decreases until it reaches that of the background at a radius of r = 455 kpc. We then calculate our final ICL fractions by measuring the total flux of the ICL and the total cluster light within this region. We note here that this limit is likely nonphysical, but it is highly dependent on the depth and the size of our observations, as proved by recent simulations (Deason et al. 2021). In contrast to measuring the ICL in fixed apertures, the total ICL fractions are independent of the point spread function (PSF) and have the advantage of diluting the possible systematic errors inherent to the calibration and reduction of the images given the larger areas involved in the calculations (Jiménez-Teja et al. 2021). The error in the ICL fraction was estimated considering three different contributions: photometric, geometrical, and cluster membership errors. The photometric error is the error inherent to the process of measuring the flux of a source with noise, and it depends on the flux measured, the gain, the rms of the sky, and the area covered by the source. The geometrical error is the error made by CICLE in the disentanglement of the BCG from the ICL. It accounts for (1) the error in the determination of the radius at which the transition from the BCG- to the ICL-dominated regions occurs, and (2) the flux missed from the BCG outer stellar halo that extends throughout the ICL-dominated region (see Sect. 2.2). This error depends on how different the profiles of the BCG and the ICL are (i.e., their magnitudes and effective radii), and it is estimated using simulated images that mimic the geometrical configuration of the BCG and the ICL. Finally, we computed the error associated with the photometric cluster membership, which depends on the completeness yielded by the ML algorithm. From the resulting completeness of 93.5% (see Sect. 2.3), we estimated the flux missed from the missing cluster galaxies in each filter and propagated it to the ICL fraction.

thumbnail Fig. 3.

Radial profiles of the total cluster, the ICL, and the background in the F160W filter. As the background was negative in some regions, we added an arbitrary quantity to the three profiles to plot the y-axis in logarithmic scale. The dashed vertical line indicates the limit of the ICL.

The final ICL fractions computed for SPT0615 are high. They range between 16.6 and 33.3%, as listed in Table 1. The final errors, which range between 0.9 and 5.0%, are also reported, along with the breakdown of the total error budget into the three error sources described above.

3. Dynamical stage of SPT0615

A clear correlation exists between the ICL fraction and the dynamical stage of the cluster (Jiménez-Teja et al. 2018). We therefore analyzed the available X-ray data and compared our results with those of previous works (e.g., Planck Collaboration XXVI 2011; Bartalucci et al. 2017; Bulbul et al. 2019; Yuan & Han 2020).

3.1. X-ray data and analysis

All archived Chandra data available for SPT0615 were used. There are 12 separate observations of SPT0615 (OBS ID: 14017, 14018, 14349, 14350, 14351, 14437, 15572, 15574, 15579, 15582, 15588 and 15589) that were taken with the Advanced CCD Imaging Spectrometer (ACIS-I), from September 15 to November 24, 2012, aimed at RA = 6h15m52 . s $ {{\overset{\text{s}}{.}}} $0, Dec = −57°46′51 . ″ $ {{\overset{\prime\prime}{.}}} $6 (PI: Mazzotta). We used CIAO 4.13 to run the standard recommended data-processing steps, reprocessing all the data with the script chandra_repro, which creates new bad-pixel files and new level-2 event files. The total effective exposure (after cleaning) is 241 ks. We created a merged event file by reprojecting all observations to a common tangent point using the reproject_obs tool. This was used solely to remove point sources and for the basic image analysis after refilling the extracted point sources using the tool dmfilth. For the spectral analysis, we used the individual event files to create separate data products and analyzed them simultaneously, following the CIAO guidelines3.

The tool specextract was used to produce spectral files and responses for the regions we analyzed. The spectra were grouped to have from 5 to 20 cnt/channel depending on the region configuration that we analyzed. Because the source angular extension is small, we used two local background regions away from the cluster center within the same CCD. One region is about 1.4 Mpc and the other is 1.9 Mpc from the cluster center at different orientations with respect to the N–S cluster X-ray elongation, to test, among other things, how strongly the cluster is contaminated near R200. Spectral fittings were carried out in XSPEC 12.11.0m using the absorbed collisional ionization equilibrium model tbabs*apec with the redshift 0.972 (Connor et al. 2019) fixed at the nominal value of the BCG and the hydrogen column density (nH) of 3.2 × 1020 cm−2 from the HI4PI Map (HI4PI Collaboration 2016) through the HEASARC nH tool4. The abundances listed here are given with respect to the photospheric value (Anders & Grevesse 1989). For the 2D image fitting and to produce residual emission, we used Sherpa v.1 (Freeman et al. 2001) in Ciao 4.1.4.

3.2. X-ray results and temperature distribution

The Chandra image of SPT0615 shows a clear X-ray elongation along a north-south direction, where the main axis has a position angle of ∼22° (Fig. 4 top right). A similar elongation is also seen in the critical curves of the gravitational lensing (Paterno-Mahler et al. 2018) and in the ICL distribution (Fig. 1). Galaxy members in the cluster tend to be distributed along this direction as well (Fig. 4 top left). The X-ray surface brightness contours also suggest some asymmetric excess emission to the east and possibly in the southern regions (Fig. 4 bottom; Sect. 3.3). At small scales, near the center, two X-ray brightest peaks are apparently configured along the N–S direction. They are displaced from the BCG by 20 kpc and 16 kpc (Fig. 5).

thumbnail Fig. 4.

Optical and X-ray images of SPT0615. Top left: HST/ACS image in the F814W filter of SPT0615. Spectroscopically confirmed member galaxies are indicated by green circles. Top right: Chandra image of the same region. The alignment seen in the galaxy distribution coincides with the major axis of the surface brightness distribution in X-rays at the inclination of ∼22° shown in the bottom figure. Bottom: X-ray isocontours of SPT0615, which show a similar elongation as the galaxy distribution and critical lensing curves (Fig. 6 of Paterno-Mahler et al. 2018). Excess X-ray emission can be seen clearly east and south, and it is marked by green arrows. North is up. East is left.

thumbnail Fig. 5.

Very central region of SPT0615. Left: Chandra X-ray image and isocontours in the very core of SPT0615. The position of the BCG is shown by a circle, and the approximate distance from each of the bright X-ray cores is plotted as well. Right: HST image of the same region. The X-ray contours are overlaid. None of the X-ray brightest central regions coincide with the BCG or any other galaxy.

We extracted the radial distribution of the intracluster gas temperatures (TX) and metal abundances (AX) using circular annuli with the same (or as close as possible) binning as was used by Bartalucci et al. (2017) using local backgrounds (Fig. 6 top). We show the results in the bottom panel of Fig. 6, where the projected TX and AX are shown in gray, and in red we show the same, but binning the two outermost annuli and the two innermost annuli because they had similar projected temperatures, to improve statistics (see Table 2). There are significant departures from isothermality within 100 kpc−200 kpc from the cluster center. To reduce the contamination from the external layers, we carried out a standard deprojection, following a simple onion-peeling-like procedure with fixed outer layer parameters and normalizations corrected for the region area. The results are shown in Table 3 and are plotted in the bottom panel of Fig. 6. The deprojection increased the significance of the temperature gradient in the 100−200 kpc region and also showed the cooler central region5 more clearly (Fig. 6 bottom in green and black for the two background choices). The region 88−128 kpc (hot ring) exhibits extremely high temperatures (TX > 15 keV), regardless of the background choice. Characteristic values correspond to shock regions in very strong mergers such as the Bullet cluster (Markevitch et al. 2002), the Pandora cluster (Merten et al. 2011), or Abell 754 (Inoue et al. 2016). The temperature of the hot ring as taken from deprojection is likely to be overestimated because we assumed isotropy for the radial temperature gradient around it. This could also make the cooler central region even cooler as well. This is a limitation of the deprojection technique for clusters with significant substructures.

thumbnail Fig. 6.

Intracluster gas temperature and radial distributions of the metal abundance. Top: Chandra ACIS-S3 image of SPT0615 with the concentric annuli (white) regions used in Bartalucci et al. (2017) and reanalyzed in this work. We also show two different background regions (B1 and B2) in this analysis (yellow) to illustrate the possible contamination of the cluster emission at different distances from the center. Bottom: projected radial profile of the temperatures and metal abundances corresponding to the concentric annuli shown in the top panel are shown in gray. We also show the same distribution where we join the inner two bins and the outer two bins in red. In green and black, we show the deprojected profiles using the B1 and B2 background regions, respectively. North is up. East is left.

Table 2.

Best-fit parameters for the projected spectral simultaneous fittings for all 12 observations.

Table 3.

Best-fit parameters for the deprojected spectral simultaneous fittings for all 12 observations.

Although the overall results are somewhat consistent with those of Bartalucci et al. (2017), the disagreement near the observed hot ring is significant. The exact reason for this cannot be immediately assessed through the radial distribution. However, the presence of this significant temperature substructure strongly suggests that the system is in the process of or has recently passed through merging and that a purely radial analysis may not be ideal to provide information about the current dynamical state of the cluster. Because of the indications of substructures throughout the cluster, we produced an adaptively smoothed temperature map by performing spectral fittings of circular regions with 3000 counts per region at least, which was found to be the minimum value, within which we could obtain good enough precision to have the discriminatory power, given the high cluster temperatures. The subsequent gridding method we used is based on an interpolation method that calculates a new value for each cell in the regular output matrix for the parameter of interest (TX) from the values of the points in the adjoining cells that are included within a given search radius. The interpolation we used follows the Kriging method and has been shown to successfully enhance the mapping of features of the parameters of interest in the intracluster medium (e.g., Dupke et al. 2007a,b). For the sake of illustration, we show in the bottom panel of Fig. 7 the smoothing overlap of the characteristic size of regions with radii smaller than 25″.

thumbnail Fig. 7.

Adaptive-smoothing temperature map of SPT0615. Top: adaptively smoothed image of the projected intracluster gas temperature in keV. The surface brightness contours in green are the same as in the bottom panel of Fig. 4. The concentric annuli used for radial spectral extraction in Fig. 6 are shown in yellow, and the annulus corresponding to the temperature spike is plotted in bold. The magenta contours show the fractional error of the temperatures corresponding to 10%, 15%, 25%, and 35% from the inside outward. In white we indicated the position of a candidate cold front from a previous merger (Sect. 3.3). Bottom: smoothing overlap of the characteristic size of regions with radii below 25″ prior to Kriging interpolation. The magenta contours correspond to the same fractional temperature errors as plotted in the top panel. North is up. East is left.

The resulting temperature map is shown in the top panel of Fig. 7. The contours shown in magenta represent different levels of significance based on the fractional temperature error. Outside the 25%−35% contours, the fractional errors grow very fast, so that the values are not informative and can be ignored. However, within the 25% confidence region lie significant substructures. Most notable of these is a hot (TX ∼ 13 keV) clump east of the main central BCG, hereafter called hot clump, which is encompassed by the hot-ring annulus, as discussed previously. The hot clump is very near the eastern excess emission, but slightly closer to cluster core. The actual eastern excess emission seems to be in a region of significantly lower temperatures (TX < 10 kev). The hot clump is also very near the large elliptical luminous galaxy (ELG) with Δz = 0.004 from the BCG (see Fig. 8). The overall configuration is consistent with a current incoming merging cluster with some significant line-of-sight component, the main central galaxy of which is the eastern ELG, where the gas is shock-heated to the high temperatures observed in the hot clump. The eastern excess in this scenario could be attributed to the intracluster medium of the incoming cluster, possibly stripped or not yet shock-heated by the merger. We show the suggested configuration in Fig. 8. A clear anisotropy appears in the NE–SW direction at larger scales as well. The high temperature shell in the direction of the southern extension is particularly interesting. It lies still within the zone of good statistical confidence. This substructure, as we hypothesize in the next section, seems to be coincident with the sudden drop in density that is characteristic of cold fronts. This suggests that a previous recent merger may have occurred in the NE–SW direction.

thumbnail Fig. 8.

Zoom-in of the temperature map shown in Fig. 7, with X-ray contours in black and a possible projected merging configuration. The position of the BCG and the large ELG of the hypothesized incoming cluster are shown in cyan (bold) as well as the other confirmed member galaxies in the region. North is up. East is left.

3.3. Image analysis

Even though the exposure time available for this cluster is too short for a detailed image analysis of the merger details, it can provide some insights about the recent evolutionary history of this cluster. For example, the level of departure from relaxation can be determined by studying the residuals from a simple 2D beta-model image fitting. We created a model with an ellipticity of 0.23 given the strong elongation seen in the X-ray image and a position angle of 115°. The original image, model, and residuals are shown in Fig. 9, along with the X-ray isocontours originally plotted in the bottom panel of Fig. 4. The residual map (Fig. 9 bottom) shows two interesting features: The confirmation of the eastern excess, and the excess emission along the elongation axis. We did not try more complex 2D distribution fittings here because the photon counts and our knowledge on the underlying density distribution are insufficient, given the merging stage of the cluster. However, this excess might be associated with a previous near plane-of-the-sky merger along this elongation axis, where the incoming system ICM had been stripped or deposited along that path (e.g., Ascasibar & Markevitch 2006).

thumbnail Fig. 9.

Residual X-ray emission of SPT0615. Top left: Chandra image of SPT0615 used for the 2D fitting with a beta model. We also indicate the angular region we used to measure the surface brightness profile in Fig. 10. Top right: beta model and best-fit parameters. Bottom: residual from the fit. The region near the hot core is very clearly detached from the other excess over the beta-model fit region. A larger excess region over a significant part of the central cluster elongation is visible as well. It extends 280 kpc to each side from the center. North is up, and east is left.

Because of the apparent extended emission toward the SW, we extracted the surface brightness profile in an angular slice (270° −330°) region in this direction (Fig. 9 top left) and show the results in Fig. 10. The surface brightness distribution marginally significantly drops at about 28″ from the center, and some extended excess emission is visible past 40″. The drop in surface brightness coincides with a rise in temperature of the gas in the hot shell by about 20%, as shown in the top panel of Fig. 7. These characteristics, if confirmed, are consistent with the presence of a cold front (Vikhlinin et al. 2001) due to a previous merger along the elongation axis of the cluster and might be associated with the excess emission in this direction.

thumbnail Fig. 10.

Surface brightness profile throughout the angular region shown in the left panel of Fig. 9, in the direction of the southern hot shell seen in the temperature map in the top panel of Fig. 7. A marginal sudden surface brightness decrease is seen about 28″ from the cluster center (framed by the two vertical red lines). This coincides with the inner hot shell. The extended emission past this point is also apparent.

4. Discussion

Jiménez-Teja et al. (2018) measured the ICL fractions in three different HST optical filters for a sample of massive clusters whose dynamical stages were well-defined by different indicators (X-ray morphology, velocity distribution, tidal features and radio structures, and others). Jiménez-Teja et al. (2019, 2021) and de Oliveira et al. (2022) expanded the sample by adding a few more clusters and confirmed that the distribution of ICL fractions measured at different rest-frame optical wavelengths differ from active (merging) to passive (relaxed) systems. Clusters in this sample spanned the redshift interval 0.02 < z < 0.56, which is significantly lower than the redshift of SPT0615, z = 0.97. However, as the two-phase scenario of ICL formation is widely accepted (i.e., the ICL is intimately linked to the BCG build-up at z > 1, while other mechanisms play a major role for z < 1), a similar distribution would be expected for the ICL fractions of SPT0615 and the intermediate-redshift clusters. For the sake of comparison, we plot in Fig. 11 the ICL fractions of SPT0615 (black) and those of the merging (red) and relaxed (blue) clusters of the previous samples. Relaxed systems have nearly constant and low fractions in the rest-frame optical wavelengths, which indicates that the primary sources of ICL are steady (e.g., tidal stripping of member galaxies as they orbit toward the center of the gravitational potential of the cluster, enhanced by dynamical friction). Moreover, constant fractions indicate a similar stellar composition of the ICL and the cluster galaxies. Conversely, merging systems display higher ICL fractions and a distinctive peak between 3800 and 4800 Å, which is explained by an excess of A- to F-type stars (as compared with the stellar composition of the galaxies) that are thrown into the ICL in a short time in a violent regime. The probable channels of these ICL stars are related to the merging stage of the cluster (e.g., tidal stripping of infalling galaxies or preprocessing in groups that are being accreted). We show here that the SPT0615 ICL fractions do not show a trend that is consistent with a passive cluster (see Fig. 11).

thumbnail Fig. 11.

ICL fractions of SPT0615 (black squares) as derived by CICLE, compared with a subsample of relaxed (blue) and merging (red) clusters at intermediate redshift. The plot is divided into six regions according to the emission peaks of the different stellar types, indicated at the top with gray letters. The red and blue regions indicate the error-weighted average of the measurements within each stellar region for the merging and relaxed systems, respectively. At the bottom, we plot the transmission curves of the filters we used for the analysis of SPT0615. All points are in the rest frame.

The relaxed state has repeatedly been claimed, however (Planck Collaboration XXVI 2011; Bartalucci et al. 2017; Morandi et al. 2015; Yuan & Han 2020). For example, Planck Collaboration XXVI (2011) confirmed the detection of cluster SPT0615 using a shallow XMM-Newton image, and classified it as relaxed based on its density profile, the nearly symmetric distribution of its hot gas, and the offset between the BCG and the X-ray peak. Bartalucci et al. (2017) previously analyzed Chandra data of SPT0615 and also argued that the cluster is relaxed, based mostly on an inferred smooth radial temperature profile and the presence of a cool core. However, their measurements of the centroid shift ⟨ω⟩, defined as the standard deviation of the projected distance between the X-ray peak and the centroid, yield contradictory results depending on the resolution of the X-ray data, that is, Chandra versus XMM-Newton. Furthermore, as we illustrated above, ⟨ω⟩ defined from an average may not be representative of the real shift given the structure near the core. Any of the shifts from X-ray peaks to BCG described here would place the cluster out of the relaxed category. The arguments based on the flux concentration for this system with XMM are too uncertain to estimate the dynamical state of the cluster (Bartalucci et al. 2019). Both Morandi et al. (2015) and Bartalucci et al. (2017) found a cool core using deeper Chandra data, while Bulbul et al. (2019) used the XMM-Newton observation and measured a higher temperature for SPT0615 when the core was included than that found by excising it, which is consistent with a lack of a cool core. Interestingly, the core-excised region analyzed by Bulbul et al. (2019) is large enough to exclude the hot clump we found here, so that their results are compatible with ours. Additionally, Bartalucci et al. (2017) found complex substructures in the cluster core with the Chandra data (not observed with XMM-Newton), including a small low surface brightness emission west of the BCG and the X-ray peaks and a horseshoe-like high surface brightness emission surrounding both peaks. More recently, Yuan & Han (2020) measured several morphological parameters using Chandra, finding inconclusive results about the dynamical stage of SPT0615. With a radically different approach, Connor et al. (2019) performed a kinematic analysis of SPT0615 with spectroscopic redshifts, finding evidence for a nonrelaxed dynamical state. They found indications for a non-Gaussian velocity dispersion of the SPT0615 galaxies that increases with clustercentric radius, along with an offset between the BCG and the remaining velocity dispersion of the cluster.

The ICL fraction values we measured for SPT0615 are more consistent with those of merging clusters, although they are slightly higher than expected in the optical and are abnormally high in the bluest band (see Fig. 11). At this point, we consider three hypotheses: (1) A strong contamination of the measured ICL fractions by bright stars, outer stellar halos of bright galaxies, and/or an incorrect estimate of the background. (2) A misclassification of this cluster, which is traditionally thought to be relaxed. Finally, and alternatively, (3) a possible second structure in the line of sight, as suggested by Paterno-Mahler et al. (2018). We rule out the first hypothesis because we thoroughly masked out all the pixels that might contain flux from the nearby bright stars, and we accurately fit the galaxies with CHEFs, inspected the residuals after removing them, and estimated and subtracted the background using NoiseChisel (see Sect. 2), which has been proved to provide excellent results in this matter (Borlaff et al. 2019; Haigh et al. 2021; Kelvin et al. 2023). As described in Sect. 2.4, the impact of these sources of contamination is included in the error budget, so it is not large enough to explain the excess found in the ICL fractions.

Our detailed X-ray analysis described in Sect. 3.2 strongly favors the unrelaxed scenario for SPT0615. This reanalysis of the Chandra data shows that the cluster is in the process of merging, possibly with multiple components. The most notable component is related to the hot clump about 150 kpc east of the center, with temperatures characteristic of shocks seen in violent mergers, and near an ELG, with a radial velocity difference of ∼600 km s−1 with respect to the BCG of the system. The temperature distribution and the galaxy distribution in the NE direction suggest that this is a pre-core-crossing merger with a plane-of-the-sky component in the E–S direction, as illustrated in Fig. 8, where the hot clump would be a result of shock-heated gas in between the two colliding cores (e.g., Ascasibar & Markevitch 2006; Bourdin et al. 2013). The large-scale temperature gradient seen along the main X-ray and galaxies axis is significant and suggests a secondary merging event along the elongation axis of the cluster, although current observations are still too shallow to determine a more precise spatial configuration.

To further investigate the third hypothesis, the presence of a group at z ∼ 0.44 with a partial projected area overlap with SPT0615, we plot in Fig. 12 the cluster members (z = 0.97) in red, the group members (z ∼ 0.44) in yellow, the ICL contours in the F160W band in blue, and the X-ray isocontours in green. The foreground group members are distributed throughout the region of SPT0615, with a certain preference south of the z = 0.97 BCG (white diamond). Several of the brightest foreground group members are located close to or within the area in which we found an excess in the gas temperature (southwestern extension and the hot shell shown in the top panel of Fig. 7). However, as several cluster members are also identified in this location, we cannot effectively determine whether this hot region is due to the possible presence of a foreground group or to the cluster itself. Furthermore, the galaxy distribution near the southwestern excess region at z ∼ 0.44 kpc seems to be highly dispersed over its encompassed ∼230 kpc, with no sign of a concentration toward any putative core. It would also be unlikely that the potentially foreground group would have similar high intracluster gas temperatures as were found in SPT0615. As for the excess found both in the X-ray isocontours and the smoothed gas temperature map, east of the BCG (see the bottom panel of Fig. 4 and the top panel of Fig. 7), the ICL contours seem to have an extended distribution or blob that matches the position of the putative incoming cluster. As no foreground group members are identified within this region, which instead encompasses three bright cluster member galaxies, we conclude that this hot-gas excess is only due to dynamical activity at the SPT0615 redshift. This corroborates its merging stage.

thumbnail Fig. 12.

Comparison of the ICL, X-ray, and double-membership analyses. The contours of the ICL in the F160W are plotted in blue, and those of the hot-gas distribution are shown in green, as in previous figures. The red circles mark the position of the galaxies that are identified as SPT0615 members, and yellow circles indicate the members of the foreground structure at z ∼ 0.44 according to our ML algorithm. The SPT0615 BCG and two of the three brightest galaxies of the foreground structure are marked with additional white and cyan diamonds, respectively. The third brightest galaxy at z ∼ 0.44 is beyond the F160W footprint, at ∼654 kpc (at z = 0.44) away from the other two, to the southwest. North is up, and east is left.

Because of the difference in richness, redshift, and the angular offset between the cluster and the possible foreground group, it is very likely that the ICL fractions are primarily dominated by SPT0615. However, some contamination from overlapping regions from the putative foreground group cannot be ruled out with the current data. Our previous works showed the importance of measuring total ICL fractions (i.e., total photometry as opposed to aperture photometry), especially to be able to use them as reliable indicators of the dynamical stage of the cluster (e.g., Jiménez-Teja et al. 2018, 2021; Joo & Jee 2023). As a consequence, partial ICL fractions or ICL fractions measured in certain regions may not be fully representative of the total ICL fractions. This is particularly true in the case of merging clusters, where the spatial distribution of ICL is often asymmetric and clumpy, which is the case of SPT0615. Taking all these considerations with caution, we measured the ICL fractions in the (apparently) less polluted region of the ICL, which is north of the BCG. In this region, only two members of the putative foreground group are identified. Additionally, this region contains the main peak of ICL, and it therefore should enclose a significant fraction of its total budget. The ICL fractions measured in this region are 28.61 ± 6.36% (F606W), 24.2 ± 2.7% (F814W), 20.8 ± 1.9% (F105W), 22.2 ± 1.1% (F125W), 18.4 ± 3.4% (F140W), and 18.9 ± 1.3% (F160W). These partial ICL fractions show (1) a similar trend as the total ICL fractions (Table 1 and Fig. 11), thus proving that SPT0615 indeed dominates the global values, (2) even more consistency with a merging state (the slight excess in the rest-frame optical F814W and F105W ICL fractions disappears), and (3) still an abnormally high ICL fraction in the F606W band.

This test suggests that the unusually high value found in the rest frame in the near-UV may indeed have a physical origin. We can speculate that this component might come from stripped early-type stars that were formed very recently during the currently ongoing merger. An upper age limit for the current merger can potentially be set by the half-life of the stars that mainly contribute in the near UV. If the UV excess is primarily formed by A0-type main-sequence stars, this would imply an upper limit of ∼0.5−0.6 Gyr, which is their typical half-life. The typical color of A0 stars is B − V = 0, and this color is equivalent to that seen in late-type galaxies (Sd), for instance, outer spiral disks and irregulars. Thus, if not much preprocessing has operated in the infalling halos, blue galaxies might be found to still enter the cluster potential and to still inject these stars into the ICL. Additionally, the star formation-density relation observed in the local Universe is reversed at z ∼ 1: The fraction of star-forming galaxies in dense environments is significantly higher than in nearby clusters, which are primarily dominated by red evolved galaxies (Elbaz et al. 2007). This would imply that the stars that are stripped from these star-forming galaxies into the ICL are bluer than those of low- and intermediate-redshift clusters (all clusters studied in previous works spanned the interval 0.18 < z < 0.56, while SPT0615 is at z = 0.97; see Fig. 11).

5. Conclusions

Despite being a high-redshift cluster, SPT0615, has previously been considered to be relaxed. Here, we revisited this scenario using two completely independent and reliable indicators of the dynamical stage: the ICL fraction, and a revised analysis of the hot-gas distribution. Previous work showed that the ICL fraction measured at different optical and infrared wavelengths exhibits a unique marker that differentiated active (merging) from relaxed (passive) systems (Jiménez-Teja et al. 2018, 2019, 2021; de Oliveira et al. 2022; Dupke et al. 2022). This has been verified in many massive clusters in low- and intermediate-redshift (0.02 < z < 0.56) systems. The merging signature was identified as a characteristic excess in the rest-frame ICL fractions measured roughly in the peak wavelength corresponding to stars of A to F spectral types, that is, between 3800 and 4800 Å. For SPT0615, we found that the ICL fraction distribution does not follow that typical of relaxed clusters, but it is more consistent with the distribution in an active system. It even shows indications of excess in the UV ICL fraction, which cannot be explained with a merger alone. It requires the radial gradient of an average spectral type observed in nearby galaxies to include very early-type stars at large radii.

We also performed an independent analysis of the existing X-ray data and noted significant departures from the expected temperature distribution for relaxed cool-core clusters, with near-core intracluster gas temperatures corresponding to those of very strong mergers. Subsequently, we produced a full temperature map that allowed us to detect several substructures (at least two major ones) that are consistent with the presence of multiple mergers, one currently in a pre-core-crossing stage, which created a shock-heated hot clump between the BCGs, and the other likely to be the remnant of a previous merger.

Because the photometric redshift distribution of the galaxies in the region of SPT0615 showed two main peaks, one at z ∼ 0.97 corresponding to the cluster and another one at z ∼ 0.44, we performed a double cluster membership analysis using a ML algorithm. The results show two clear red sequences with 176 and 36 members, respectively. The projected distribution of the cluster and foreground group galaxies overlap partially, with a higher concentration of group members southwest of the cluster BCG. However, the ICL analysis performed in regions with little or no projected overlap does not indicate a significant change in the overall results. Furthermore, the lack of a concentration around some putative foreground BCG by either galaxies or intracluster gas and the general spread of the foreground galaxies suggest that there is no significant contamination. We list the main results we found from the HST and X-ray analyses below.

  • The ICL contours show an extended ICL that is elongated along an axis with a position angle of ∼30°, and two main clumps north and southwest of the BCG.

  • The ICL fractions range from 16.6 to 36.3%, with a spectral distribution consistent with that of merging clusters, but it is slightly higher than expected in the rest-frame optical bands.

  • The ICL fraction in the bluest filter, F606W, is abnormally high when compared with that of low- and intermediate-redshift clusters, either merging or relaxed, but it is consistent with an extension of the ICL fraction peak seen in merging clusters. This would be consistent with observing a cluster merger with a upper limit age of ∼0.5 Gyr if the injected ICL stars are of very early type, A0 or earlier. This would imply that late-type galaxies were not strongly pre-processed before they infall into the cluster potential. The reversal of the star formation-density relation in the distant Universe can also cause more bluer stars in the ICL.

  • The X-ray surface brightness shows an elongation along a position angle of ∼22°, the same as seen in the optical analysis. Significant brightness excess with respect to a standard 2D beta profile is observed in this elongation.

  • The deprojected temperature profile confirms a cooler central region, but shows a strong temperature fluctuations within 100−200 kpc.

  • The adaptively smoothed temperature map shows significant azimuthal substructures, in particular, a hot clump about 150 kpc far from the cluster center near an ELG. This is consistent with a strong ongoing (pre-core-crossing) merger with temperatures consistent with shock-heated gas found in other strong mergers. This clump traces an ICL filament well that is associated with three bright cluster member galaxies. This corroborates the ongoing merging hypothesis.

  • There is significant evidence of other temperature substructures that is consistent with previous or ongoing secondary mergers. In particular, the surface brightness profile southwest of the elongation axis of the cluster seems to have a slight drop at ∼28″ from the center, right before a abrupt temperature rise. This is characteristic of cold fronts of the merging (not sloshing) type (Dupke et al. 2007b; Markevitch & Vikhlinin 2007). If this is confirmed by future observations, it would indicate that a previous merger occurred recently along the elongation axis and that the gaseous core of the incoming system would be ∼280 kpc southwest of the cluster center.

This work highlights the power of the ICL, and in particular, of the ICL fractions, in determining the dynamical stage of clusters. CICLE provides unbiased measurements of the ICL fractions that are able to discover anomalies that can be associated with merging activity and/or projection effects, in particular, when coupled with X-ray observations with high spatial resolution, such as those of Chandra. This is particularly important for studying the evolution of high-z clusters, given their small angular sizes. The forthcoming JWST observations6 of SPT0615 will provide crucial insights for a determination of the merging configuration.

5

We are not convinced that this cooler region constitutes a classic cool core, especially given the temperature distribution shown in Fig. 7.

Acknowledgments

All authors sincerely thank the anonymous referee for his/her useful and kind comments, which have certainly improved the quality of this manuscript. Y.J.-T. acknowledges financial support from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 898633 and the MSCA IF Extensions Program of the Spanish National Research Council (CSIC). R.A.D. acknowledges partial support from the CNPq grants 308105/2018-4 & 312565/2022-4. J.M.V. thanks support from project PID2019-107408GB-C44 (Spanish Ministerio de Ciencia e Innovación). Y.J.-T., R.A.D., and J.M.V. acknowledge support from the State Agency for Research of the Spanish MCIU through the “Center of Excellence Severo Ochoa” award to the Instituto de Astrofísica de Andalucía (SEV-2017-0709) and grant CEX2021-001131-S funded by MCIN/AEI/10.13039/501100011033. P.A.A.L. thanks the support of CNPq (grants 433938/2018-8 e 312460/2021-0) and FAPERJ (grant E-26/200.545/2023). We thank Dr. Rebeca Batalha for helpful discussions. This work is based on observations taken by the RELICS Treasury Program (GO 14096) with the NASA/ESA HST, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555.

References

  1. Adami, C., Mazure, A., Biviano, A., Katgert, P., & Rhee, G. 1998, A&A, 331, 493 [NASA ADS] [Google Scholar]
  2. Adami, C., Durret, F., Guennou, L., & Da Rocha, C. 2013, A&A, 551, A20 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Akhlaghi, M. 2019, arXiv e-prints [arXiv:1909.11230] [Google Scholar]
  4. Akhlaghi, M., & Ichikawa, T. 2015, ApJS, 220, 1 [Google Scholar]
  5. Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197 [Google Scholar]
  6. Ascasibar, Y., & Markevitch, M. 2006, ApJ, 650, 102 [Google Scholar]
  7. Bartalucci, I., Arnaud, M., Pratt, G. W., et al. 2017, A&A, 598, A61 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  8. Bartalucci, I., Arnaud, M., Pratt, G. W., Démoclès, J., & Lovisari, L. 2019, A&A, 628, A86 [EDP Sciences] [Google Scholar]
  9. Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  10. Borlaff, A., Trujillo, I., Román, J., et al. 2019, A&A, 621, A133 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  11. Bourdin, H., Mazzotta, P., Markevitch, M., Giacintucci, S., & Brunetti, G. 2013, ApJ, 764, 82 [Google Scholar]
  12. Bulbul, E., Chiu, I. N., Mohr, J. J., et al. 2019, ApJ, 871, 50 [Google Scholar]
  13. Burke, C., Collins, C. A., Stott, J. P., & Hilton, M. 2012, MNRAS, 425, 2058 [Google Scholar]
  14. Coe, D., Salmon, B., Bradač, M., et al. 2019, ApJ, 884, 85 [Google Scholar]
  15. Connor, T., Kelson, D. D., Blanc, G. A., & Boutsia, K. 2019, ApJ, 878, 66 [NASA ADS] [CrossRef] [Google Scholar]
  16. Conroy, C., Wechsler, R. H., & Kravtsov, A. V. 2007, ApJ, 668, 826 [NASA ADS] [CrossRef] [Google Scholar]
  17. de Oliveira, N. O. L., Jiménez-Teja, Y., & Dupke, R. 2022, MNRAS, 512, 1916 [NASA ADS] [CrossRef] [Google Scholar]
  18. Deason, A. J., Oman, K. A., Fattahi, A., et al. 2021, MNRAS, 500, 4181 [Google Scholar]
  19. DeMaio, T., Gonzalez, A. H., Zabludoff, A., Zaritsky, D., & Bradač, M. 2015, MNRAS, 448, 1162 [NASA ADS] [CrossRef] [Google Scholar]
  20. DeMaio, T., Gonzalez, A. H., Zabludoff, A., et al. 2018, MNRAS, 474, 3009 [Google Scholar]
  21. Dupke, R. A., Mirabal, N., Bregman, J. N., & Evrard, A. E. 2007a, ApJ, 668, 781 [Google Scholar]
  22. Dupke, R., White, R. E., III, & Bregman, J. N. 2007b, ApJ, 671, 181 [NASA ADS] [CrossRef] [Google Scholar]
  23. Dupke, R. A., Jimenez-Teja, Y., Su, Y., et al. 2022, ApJ, 936, 59 [NASA ADS] [CrossRef] [Google Scholar]
  24. Elbaz, D., Daddi, E., Le Borgne, D., et al. 2007, A&A, 468, 33 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  25. Fadda, D., Girardi, M., Giuricin, G., Mardirossian, F., & Mezzetti, M. 1996, ApJ, 473, 670 [NASA ADS] [CrossRef] [Google Scholar]
  26. Freeman, P., Doe, S., & Siemiginowska, A. 2001, SPIE Conf. Ser., 4477, 76 [NASA ADS] [Google Scholar]
  27. Haigh, C., Chamba, N., Venhola, A., et al. 2021, A&A, 645, A107 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  28. HI4PI Collaboration (Ben Bekhti, N., et al.) 2016, A&A, 594, A116 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  29. Inoue, S., Hayashida, K., Ueda, S., et al. 2016, PASJ, 68, S23 [NASA ADS] [CrossRef] [Google Scholar]
  30. Iodice, E., Spavone, M., Cantiello, M., et al. 2017, ApJ, 851, 75 [Google Scholar]
  31. Jiménez-Teja, Y., & Benítez, N. 2012, ApJ, 745, 150 [Google Scholar]
  32. Jiménez-Teja, Y., & Dupke, R. 2016, ApJ, 820, 49 [Google Scholar]
  33. Jiménez-Teja, Y., Dupke, R., Benítez, N., et al. 2018, ApJ, 857, 79 [Google Scholar]
  34. Jiménez-Teja, Y., Dupke, R. A., Lopes de Oliveira, R., et al. 2019, A&A, 622, A183 [Google Scholar]
  35. Jiménez-Teja, Y., Vílchez, J. M., Dupke, R. A., et al. 2021, ApJ, 922, 268 [CrossRef] [Google Scholar]
  36. Joo, H., & Jee, M. J. 2023, Nature, 613, 37 [NASA ADS] [CrossRef] [Google Scholar]
  37. Katgert, P., Mazure, A., Perea, J., et al. 1996, A&A, 310, 8 [NASA ADS] [Google Scholar]
  38. Kelvin, L. S., Hasan, I., & Tyson, J. A. 2023, MNRAS, 520, 2484 [NASA ADS] [CrossRef] [Google Scholar]
  39. Ko, J., & Jee, M. J. 2018, ApJ, 862, 95 [Google Scholar]
  40. Koekemoer, A. M. 2002, HST Dither Handbook, HST Data Handbooks [Google Scholar]
  41. Lidman, C., Suherli, J., Muzzin, A., et al. 2012, MNRAS, 427, 550 [Google Scholar]
  42. Lim, S., Côté, P., Peng, E. W., et al. 2020, ApJ, 899, 69 [CrossRef] [Google Scholar]
  43. Lopes, P. A. A., & Ribeiro, A. L. B. 2020, MNRAS, 493, 3429 [NASA ADS] [CrossRef] [Google Scholar]
  44. Lopes, P. A. A., de Carvalho, R. R., Kohl-Moreira, J. L., & Jones, C. 2009, MNRAS, 392, 135 [NASA ADS] [CrossRef] [Google Scholar]
  45. Markevitch, M., & Vikhlinin, A. 2007, Phys. Rep., 443, 1 [Google Scholar]
  46. Markevitch, M., Gonzalez, A. H., David, L., et al. 2002, ApJ, 567, L27 [Google Scholar]
  47. Melnick, J., Giraud, E., Toledo, I., Selman, F., & Quintana, H. 2012, MNRAS, 427, 850 [NASA ADS] [CrossRef] [Google Scholar]
  48. Merten, J., Coe, D., Dupke, R., et al. 2011, MNRAS, 417, 333 [NASA ADS] [CrossRef] [Google Scholar]
  49. Montes, M., & Trujillo, I. 2014, ApJ, 794, 137 [Google Scholar]
  50. Montes, M., & Trujillo, I. 2018, MNRAS, 474, 917 [Google Scholar]
  51. Morandi, A., Sun, M., Forman, W., & Jones, C. 2015, MNRAS, 450, 2261 [NASA ADS] [CrossRef] [Google Scholar]
  52. Morishita, T., Abramson, L. E., Treu, T., et al. 2017, ApJ, 846, 139 [Google Scholar]
  53. Murante, G., Giovalli, M., Gerhard, O., et al. 2007, MNRAS, 377, 2 [Google Scholar]
  54. Paterno-Mahler, R., Sharon, K., Coe, D., et al. 2018, ApJ, 863, 154 [Google Scholar]
  55. Patrikalakis, N. M., & Maekawa, T. 2010, Shape Interrogation for Computer Aided Design and Manufacturing (Springer) [Google Scholar]
  56. Peebles, P. J., & Ratra, B. 2003, Rev. Mod. Phys., 75, 559 [NASA ADS] [CrossRef] [Google Scholar]
  57. Planck Collaboration XXVI. 2011, A&A, 536, A26 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  58. Planck Collaboration XXVII. 2016, A&A, 594, A27 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  59. Puchwein, E., Springel, V., Sijacki, D., & Dolag, K. 2010, MNRAS, 406, 936 [NASA ADS] [Google Scholar]
  60. Ragusa, R., Iodice, E., Spavone, M., et al. 2023, A&A, 670, L20 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  61. Román, J., Trujillo, I., & Montes, M. 2020, A&A, 644, A42 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  62. Rudick, C. S., Mihos, J. C., & McBride, C. 2006, ApJ, 648, 936 [NASA ADS] [CrossRef] [Google Scholar]
  63. Rudick, C. S., Mihos, J. C., & McBride, C. K. 2011, ApJ, 732, 48 [Google Scholar]
  64. Venhola, A., Peletier, R., Laurikainen, E., et al. 2018, A&A, 620, A165 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  65. Venhola, A., Peletier, R. F., Salo, H., et al. 2022, A&A, 662, A43 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  66. Vikhlinin, A., Markevitch, M., & Murray, S. S. 2001, ApJ, 551, 160 [CrossRef] [Google Scholar]
  67. Williamson, R., Benson, B. A., High, F. W., et al. 2011, ApJ, 738, 139 [NASA ADS] [CrossRef] [Google Scholar]
  68. Yuan, Z. S., & Han, J. L. 2020, MNRAS, 497, 5485 [Google Scholar]

All Tables

Table 1.

Limiting surface brightness calculated in boxes of 3 × 3 arcsec2, ICL fractions computed with CICLE, and breakdown of the ICL fraction error into the three sources considered.

In the text
Table 2.

Best-fit parameters for the projected spectral simultaneous fittings for all 12 observations.

In the text
Table 3.

Best-fit parameters for the deprojected spectral simultaneous fittings for all 12 observations.

In the text

All Figures

thumbnail Fig. 1.

ICL isocontours superimposed on the original images in the six HST ACS and WFC3 filters. For each band, we plot ten logarithmically spaced isocontours. The lowest level is calculated from the detection limit of the ICL (where it converges with the background level), and the highest level corresponds to its maximum value. We report the surface brightness limits for the lowest and highest isocontours inside each panel. Two main ICL clumps appear in all bands. They are more clearly separated in the bluer filters.
In the text
thumbnail Fig. 2.

Photometric cluster membership. Top: photometric redshift distribution of the galaxies in the region of SPT0615. Bottom: color-magnitude diagram with the members of SPT0615 at z = 0.97 in red and those of the possible foreground structure at z ∼ 0.44 in black. Gray points correspond to interlopers that do not belong to any of the two structures.
In the text
thumbnail Fig. 3.

Radial profiles of the total cluster, the ICL, and the background in the F160W filter. As the background was negative in some regions, we added an arbitrary quantity to the three profiles to plot the y-axis in logarithmic scale. The dashed vertical line indicates the limit of the ICL.
In the text
thumbnail Fig. 4.

Optical and X-ray images of SPT0615. Top left: HST/ACS image in the F814W filter of SPT0615. Spectroscopically confirmed member galaxies are indicated by green circles. Top right: Chandra image of the same region. The alignment seen in the galaxy distribution coincides with the major axis of the surface brightness distribution in X-rays at the inclination of ∼22° shown in the bottom figure. Bottom: X-ray isocontours of SPT0615, which show a similar elongation as the galaxy distribution and critical lensing curves (Fig. 6 of Paterno-Mahler et al. 2018). Excess X-ray emission can be seen clearly east and south, and it is marked by green arrows. North is up. East is left.
In the text
thumbnail Fig. 5.

Very central region of SPT0615. Left: Chandra X-ray image and isocontours in the very core of SPT0615. The position of the BCG is shown by a circle, and the approximate distance from each of the bright X-ray cores is plotted as well. Right: HST image of the same region. The X-ray contours are overlaid. None of the X-ray brightest central regions coincide with the BCG or any other galaxy.
In the text
thumbnail Fig. 6.

Intracluster gas temperature and radial distributions of the metal abundance. Top: Chandra ACIS-S3 image of SPT0615 with the concentric annuli (white) regions used in Bartalucci et al. (2017) and reanalyzed in this work. We also show two different background regions (B1 and B2) in this analysis (yellow) to illustrate the possible contamination of the cluster emission at different distances from the center. Bottom: projected radial profile of the temperatures and metal abundances corresponding to the concentric annuli shown in the top panel are shown in gray. We also show the same distribution where we join the inner two bins and the outer two bins in red. In green and black, we show the deprojected profiles using the B1 and B2 background regions, respectively. North is up. East is left.
In the text
thumbnail Fig. 7.

Adaptive-smoothing temperature map of SPT0615. Top: adaptively smoothed image of the projected intracluster gas temperature in keV. The surface brightness contours in green are the same as in the bottom panel of Fig. 4. The concentric annuli used for radial spectral extraction in Fig. 6 are shown in yellow, and the annulus corresponding to the temperature spike is plotted in bold. The magenta contours show the fractional error of the temperatures corresponding to 10%, 15%, 25%, and 35% from the inside outward. In white we indicated the position of a candidate cold front from a previous merger (Sect. 3.3). Bottom: smoothing overlap of the characteristic size of regions with radii below 25″ prior to Kriging interpolation. The magenta contours correspond to the same fractional temperature errors as plotted in the top panel. North is up. East is left.
In the text
thumbnail Fig. 8.

Zoom-in of the temperature map shown in Fig. 7, with X-ray contours in black and a possible projected merging configuration. The position of the BCG and the large ELG of the hypothesized incoming cluster are shown in cyan (bold) as well as the other confirmed member galaxies in the region. North is up. East is left.
In the text
thumbnail Fig. 9.

Residual X-ray emission of SPT0615. Top left: Chandra image of SPT0615 used for the 2D fitting with a beta model. We also indicate the angular region we used to measure the surface brightness profile in Fig. 10. Top right: beta model and best-fit parameters. Bottom: residual from the fit. The region near the hot core is very clearly detached from the other excess over the beta-model fit region. A larger excess region over a significant part of the central cluster elongation is visible as well. It extends 280 kpc to each side from the center. North is up, and east is left.
In the text
thumbnail Fig. 10.

Surface brightness profile throughout the angular region shown in the left panel of Fig. 9, in the direction of the southern hot shell seen in the temperature map in the top panel of Fig. 7. A marginal sudden surface brightness decrease is seen about 28″ from the cluster center (framed by the two vertical red lines). This coincides with the inner hot shell. The extended emission past this point is also apparent.
In the text
thumbnail Fig. 11.

ICL fractions of SPT0615 (black squares) as derived by CICLE, compared with a subsample of relaxed (blue) and merging (red) clusters at intermediate redshift. The plot is divided into six regions according to the emission peaks of the different stellar types, indicated at the top with gray letters. The red and blue regions indicate the error-weighted average of the measurements within each stellar region for the merging and relaxed systems, respectively. At the bottom, we plot the transmission curves of the filters we used for the analysis of SPT0615. All points are in the rest frame.
In the text
thumbnail Fig. 12.

Comparison of the ICL, X-ray, and double-membership analyses. The contours of the ICL in the F160W are plotted in blue, and those of the hot-gas distribution are shown in green, as in previous figures. The red circles mark the position of the galaxies that are identified as SPT0615 members, and yellow circles indicate the members of the foreground structure at z ∼ 0.44 according to our ML algorithm. The SPT0615 BCG and two of the three brightest galaxies of the foreground structure are marked with additional white and cyan diamonds, respectively. The third brightest galaxy at z ∼ 0.44 is beyond the F160W footprint, at ∼654 kpc (at z = 0.44) away from the other two, to the southwest. North is up, and east is left.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.