Mon. Not. R. Astron. Soc. 000, 1–?? (0000) Printed 20 August 2015 (MN LATEX style file v2.2) The XMM-Newton view of the central degrees of the Milky Way G. Ponti1? , M. R. Morris2 , R. Terrier3 , F. Haberl1 , R. Sturm1 , M. Clavel3,4 , S. Soldi3,4 , A. Goldwurm3,4 , P. Predehl1 , K. Nandra1 , G. Belanger5 , R. S. Warwick6 and V. Tatischeff7 1 Max Planck Institut für Extraterrestrische Physik, 85748, Garching, Germany Department of Physics and Astronomy, University of California, Los Angeles, CA 90095-1547, USA 3 Unité mixte de recherche Astroparticule et Cosmologie, 10 rue Alice Domon et Léonie Duquet, 75205 Paris, France 4 Service d’Astrophysique (SAp), IRFU/DSM/CEA-Saclay, 91191 Gif-sur-Yvette Cedex, France 5 ESA/ESAC, PO Box 78, 28691 Villanueva de la Cañada, Spain 6 Department of Physics and Astronomy, University of Leicester, University Road, Leicester, LE1 7RH, UK 7 Centre de Sciences Nucléaires et de Sciences de la Matière, IN2P3-CNRS and Univ Paris-Sud, F-91405 Orsay Cedex, France arXiv:1508.04445v1 [astro-ph.HE] 18 Aug 2015 2 20 August 2015 ABSTRACT The deepest XMM-Newton mosaic map of the central 1.5◦ of the Galaxy is presented, including a total of about 1.5 Ms of EPIC-pn cleaned exposures in the central 15” and about 200 ks outside. This compendium presents broad-band X-ray continuum maps, soft X-ray intensity maps, a decomposition into spectral components and a comparison of the X-ray maps with emission at other wavelengths. Newly-discovered extended features, such as supernova remnants (SNRs), superbubbles and X-ray filaments are reported. We provide an atlas of extended features within ±1 degree of Sgr A? . We discover the presence of a coherent X-ray emitting region peaking around G0.1-0.1 and surrounded by the ring of cold, mid-IR-emitting material known from previous work as the ”Radio Arc Bubble” and with the addition of the X-ray data now appears to be a candidate superbubble. Sgr A’s bipolar lobes show sharp edges, suggesting that they could be the remnant, collimated by the circumnuclear disc, of a SN explosion that created the recently discovered magnetar, SGR J1745-2900. Soft X-ray features, most probably from SNRs, are observed to fill holes in the dust distribution, and to indicate a direct interaction between SN explosions and Galactic center (GC) molecular clouds. We also discover warm plasma at high Galactic latitude, showing a sharp edge to its distribution that correlates with the location of known radio/mid-IR features such as the ”GC Lobe”. These features might be associated with an inhomogeneous hot ”atmosphere” over the GC, perhaps fed by continuous or episodic outflows of mass and energy from the GC region. Key words: Galaxy: centre; nucleus; interstellar medium; ISM: supernova remnants; bubbles; kinematics and dynamics; X-rays: binaries; diffuse background; ISM; plasmas; methods: data analysis; 1 INTRODUCTION At a distance of only ∼ 8 kpc, the center of the Milky Way is the closest Galactic nucleus, allowing us to directly image, with incomparable spatial resolution, the physical processes typical of galactic nuclei. The central region of the Galaxy is one of the richest laboratories for astrophysics (Genzel et al. 2010; Morris et al. 2012; Ponti et al. 2013). Within the inner ∼ 200 pc about 3 − 5 × 107 M of molecular material are concentrated, the so called Central Molecular Zone (CMZ). This corresponds to about 1 % of the molecular mass of the entire Galaxy and it is concentrated in a region of about ∼ 10−6 of its volume (Morris & Serabyn 1996). In this region ? [email protected] c 0000 RAS many thousands of persistent and transient point-like X-ray sources are embedded, such as active stars, bright accreting binary systems (and many more quiescent massive bodies) and cataclysmic variables, which have been beautifully imaged thanks to the superior spatial resolution of Chandra (Wang et al. 2002; Muno et al. 2003; 2009). One of the best jewels in the GC is Sgr A? , the electromagnetic counterpart of the closest supermassive black hole (BH; Genzel et al. 2010). In addition to this large population of point sources, extended X-ray sources, such as supernova remnants, non-thermal filaments, pulsar wind nebulae, and massive star clusters, populate the GC (Wang et al. 2002). The GC is considered a mini-starburst environment, giving us the possibility to study the interaction between supernova remnants (SNRs) and molecular clouds and the impact of massive-and-young star clusters on their surroundings. It 2 G. Ponti et al. allows us to image, in superb detail, the creation and evolution of bubbles and superbubbles and the generation of Galactic outflows, powered by past starbursts and/or accretion events onto Sgr A? , and their impact on the GC environment. Warm (kT ∼ 1 keV) and hot (kT ∼ 6.5 keV) thermal plasma emission plus non-thermal hard X-ray emission associated with Xray reflection nebulae (see Ponti et al. 2013 for a review) pervade the central region, producing a high background of soft and hard Xray radiation. About 90% (Ebisawa et al. 2001; Wang et al. 2002) of the soft X-ray emission appears to be due to a diffuse, patchy and thermal component (Bamba et al. 2002) with a temperature kT ∼ 1 keV, most probably associated with supernova remnants. The origin of the hot component is, instead, still highly debated. At ∼ 1.5◦ from the GC, ∼ 80 % of this emission has been resolved into point sources (e.g., accreting white dwarfs and coronally active stars) by a deep Chandra observation (Revnivtsev et al. 2009). Although the intensity of the hot plasma emission increases rapidly towards the GC, point sources continue to make a substantial contribution to the observed hard emission (Muno et al. 2004; Heard & Warwick 2013a). Additionally, some of the emission may arise due to scattering of the radiation from bright X-ray binaries by the dense interstellar medium (Sunyaev et al. 1993; Molaro et al. 2014). Nevertheless, it is not excluded that a truly diffuse hot-plasma component is also present in the GC (Koyama et al. 2009; Uchiyama et al. 2013). Such hot plasma would be unbound to the Galaxy and it would require a huge energy (E ∼ 1055 erg) and energy loss rate of the mass outflow of ∼ 1043 erg s−1 , corresponding to a rate of 1 supernova/yr, to continuously replenish it (Tanaka 2002). However, it has recently been proposed that such hot plasma might be trapped by the GC magnetic field (Nishiyama et al. 2013). Indeed, the magnetic field is thought to be an important ingredient of the GC environment. The first high-resolution radio images of the Milky Way center (see bottom panel of Fig. 6), revealed the presence of many straight, long (up to ∼ 20 − 30 pc) and thin (with width < ∼0.1 pc), linearly polarised vertical filaments with spectral index consistent with synchrotron radiation (YusefZadeh et al. 1984; 1987a,b; Anantharamaiah et al. 1991; Lang et al. 1999; LaRosa et al. 2000). These filaments are hypothesized to be magnetic flux tubes trapping energetic electrons and therefore tracing the diffuse interstellar GC poloidal magnetic field (Morris & Yusef-Zadeh 1985; Lang et al. 1999). A staggeringly powerful poloidal magnetic field pervading the GC, with a field strength of B > ∼ 50 µG, and very possibly B ∼ 1 mG, has been inferred (Morris 1990; Crocker et al. 2010; Ferriere et al. 2011). The details of the physical process creating the filaments and energising the magnetic field are still debated; however, it appears clear that the magnetic filaments are interacting with the ionised surfaces of massive molecular clouds. Recent far-infrared/sub-millimeter polarization studies of thermal dust emission made it possible to probe the direction of the interstellar magnetic field inside dense molecular clouds. The magnetic field threading GC molecular clouds is measured to be parallel to the Galactic plane (Novak et al. 2003; Chuss et al. 2003; Nishiyama et al. 2009). Therefore, it appears that the large-scale GC magnetic field is poloidal in the diffuse interstellar medium and toroidal in dense regions in the plane. If the strength of the diffuse magnetic field is on the high side (B ∼ 1 mG) a huge amount of magnetic energy, E ∼ 1055 erg, would be stored in the central ∼ 300 pc. This is comparable to the kinetic energy associated with the rotation of the gas in the CMZ. Therefore, it is thought to be a key player for the GC physics and phenomenology. A large scale structure with a possible magnetic origin and appearing to be interacting with massive clouds of the CMZ (similar to the non-thermal filaments) is the Galactic center lobe (GCL). The GCL has a limb brightened shell structure in the 10.5 GHz map, defined primarily by two spurs (see Fig. 1 and 2 of Law et al. 2009). The eastern one arises from the location of the GC Radio Arc1 , while the second starts from the Sgr C thread. It was proposed that the GCL is produced by channelling of plasma from the Galactic plane, induced by energetic GC activity (e.g. episode of AGN activity, or a large mass outflow due to the high star formation rate, etc.; see Law et al. 2011) or from twisting of poloidal magnetic field lines by Galactic rotation (Sofue et al. 1984; 1985; Uchida 1985; 1990; Shibata 1989). Located at the western limb of the GCL is an interesting feature, AFGL 5376 (Uchida et al. 1994), an unusually warm, shock heated and extended IR source, thought to be associated with the GCL. All major X-ray telescopes devoted a significant fraction of their time to the study of the GC. Chandra invested several Ms to monitor both Sgr A? ’s activity (Baganoff et al. 2001; 2003; Neilsen et al. 2013) as well as diffuse soft and hard X-ray emission (Wang et al. 2002; Park et al. 2004). Suzaku and Swift also performed large observational campaigns to scan the Milky Way center (Koyama et al. 2007; Degenaar & Wijnands 2010) and monitor the transients in the region (Degenaar et al. 2012). The study of the GC is one of the key programs of the NuSTAR mission (Harrison et al. 2013; Barriere et al. 2014; Mori et al. 2013). XMM-Newton completed a first shallow (∼ 30 ks total cleaned exposure in each point) scan of the CMZ within a couple of years after launch (see the conference proceedings: Sakano et al. 2003; 2004; Decourchelle et al. 2003). A larger amount of time (more than ∼ 1.5 Ms) has been invested by XMM-Newton on studying the emission properties of Sgr A? (Goldwurm et al. 2003; Porquet et al. 2003; 2008; Belanger et al. 2005; Trap et al. 2011; Mossoux et al. 2014), focussing on the central ∼ 15 arcmin, only. Using the XMM-Newton observations from the shallow scan of the CMZ together with a number of the Sgr A? pointings, Heard & Warwick (2013a,b) have investigated the distribution of the X-ray emission within the central region of the Galaxy. With the aim of studying the propagation of echoes of the past GC activity within the CMZ (Sunyaev et al. 1993; Koyama et al. 1996; 2008; Revnivtsev et al. 2004; Muno et al. 2007; Inui et al. 2009; Ponti et al. 2010; 2013; Terrier et al. 2010; Nobukawa et al. 2011; Capelli et al. 2011; 2012; Clavel et al. 2013; 2014; Krivonos et al. 2014), recently, a new deep (with ∼ 100 ks exposure at each location) XMM-Newton scan of the CMZ has been completed (in fall 2012). We present here the combined images of both the new and old XMM-Newton scans, as well as all the XMMNewton observations within the central degree of the Galaxy. In §2 we present the data reduction process and the key steps to produce the GC EPIC mosaic maps. Section 3 introduces the broad band X-ray images, discussing the (transient) emission from the brightest point sources, the contribution from the foreground emission, as well as the soft and hard GC diffuse emission. In section 4, the narrow band images at the energies of the soft X-ray lines are displayed. Section 5 presents a new technique of spectral-imaging decomposition of the soft X-ray emission into three physical components. Section 6 presents an atlas of all the new and known diffuse features within the surveyed area. Section 7 presents the comparison with the distribution of column density of intervening mat- 1 This is a well known radio feature (see Fig. 6) composed of an array of straight, long, thin and linearly polarised vertical filaments, indicating the importance of the GC magnetic field. c 0000 RAS, MNRAS 000, 1–?? 3 ter. Discussion and conclusions are in § 8 and 9, respectively. Hereinafter, unless otherwise stated, we will state all locations and positions in Galactic coordinates and Galactic cardinal points. Errors are given at 90 % confidence for one interesting parameter. 2 DATA REDUCTION AND CLEANING The new XMM-Newton CMZ scan has been performed in 2012 starting on August 30th and ending on October 10th . It comprises of 16 XMM-Newton observations all performed with all the EPIC instruments in full-frame CCD readout mode with the medium optical blocking filters applied (we refer to Tab. 8 and 9 for more details on the instruments set-ups). This paper is not limited to the use of the 2012 XMM-Newton scan of the CMZ. Instead it is using all XMM-Newton observations pointed within 1 degree from Sgr A? . Therefore we combined the 16 observations of the new CMZ scan with the 14 observations of the previous CMZ scan accumulated between 2000 and 2002. We also include the 30 observations pointing at Sgr A? and other 49 observations aimed at studies of different sources in the vicinity of Sgr A? (see Tab. 8 and 9). We performed the analysis of the EPIC data with the version 13.0.0 of the XMM-Newton Science Analysis System (SAS). Periods of increased particle background have been removed from the data. To perform this, we first selected the Good Time Intervals (GTI) starting from the 7-15 keV background light curves, then we applied a threshold of 8 and 2.5 cts ks−1 arcmin−2 for EPIC pn and EPIC MOS, respectively (see e.g. Haberl et al. 2012). The chosen thresholds efficiently cut out all the periods of most extreme activity of soft proton flares. We noted, however, that an enhanced, but weak, background activity was still present in the data during several observations. Because of the non-uniform distribution of the GC diffuse emission, lowering the threshold uniformly in all observations, would result in cutting truly good time intervals in observations with higher GC diffuse emission. Thus we decided to visually inspect the background light curve of all data-sets and select a different threshold for each observation (see Tab. 8 and 9). Such as in Haberl et al. (2012), when the data from several detectors were available, we combined the GTIs using only common time intervals, otherwise we included GTIs of the single detector. Most of the 2012 CMZ scan data were affected by negligible particle flaring activity. On the other hand, many of the previous observations have been severely affected by soft proton flares (see the reduction in exposure in Tab. 8 and 9). To prevent infrared, optical and UV photons from bright sources in the field of view that would increase the noise and degrade the CCD energy scale, the different XMM-Newton observations have been performed with different filters applied, according to the optical-UV brightness of the sources in the field of view (see Tab. 8 and 9). In particular, we used the filter wheel closed observations to remove the internal EPIC background. 2.1 Images and exposure maps Images and exposure maps, corrected for vignetting, have been produced with an image pixel size of 2” × 2” for each energy band (for the definition of all bands, see § 2.2). To increase the sky coverage, we selected EPIC-pn events requiring (FLAG & 0xfa0000) = 0, which also includes events in pixels next to bad pixels or bad columns. Moreover, we used single to double pixel events. c 0000 RAS, MNRAS 000, 1–?? EPIC-MOS events were required to have FLAG = 0 and single to quadruple-pixel events were allowed. Figure 1 shows the combined EPIC exposure map that covers the entire CMZ. Such as done in Sturm et al. (2013), EPIC-MOS1 and -MOS2 exposures are weighted by a factor of 0.4 relative to EPIC-pn, before being added to the latter, to account for the lower effective area. Therefore, the exposure times obtained correspond to the equivalent total EPIC-pn exposure time. This allows us to obtain a better combination of EPIC-pn and EPIC-MOS data for image display purposes. We note, however, that the fluxes can not be easily read out directly from these combined images. Therefore, the line profiles and the measured fluxes/luminosities are computed from the EPIC-pn and each EPIC-MOS map separately and then combined (averaged) to obtain a better signal to noise. The top panel of Fig. 1 shows that more than 1.5 Ms of clean (after cut of time intervals during increased particle background activity) exposure time (EPIC-pn equivalent) has been accumulated on Sgr A? and over ∼ 100 − 200 ks are present in each point of the CMZ. The few pointings above and below the plane have exposures between ∼ 15 − 40 ks. Regions with less than 7.2 ks of equivalent EPIC-pn exposure have been masked out. To check the impact of the bright transients on the images and on the physical quantities under investigation, two sets of maps have been created. The first series keeps all bright transients and point sources, while the second set removes their emission by excising from the data extended regions including the transients whenever they were in outburst (see section 3.1). The middle and bottom panels of Fig. 1 show the exposure maps (computed in the same way) for the observations of new and old CMZ scans, separately. The maximum exposure times are ∼ 190 ks and ∼ 45 ks during the new and old scan, respectively. 2.2 Energy bands We created images in several energy bands (see Tab. 1). Figure 2 shows the EPIC-pn spectra of the extended emission from several regions within the CMZ. In red and black are the spectra from the G0.11-0.11 and Center Superbubble regions, respectively (see Fig. 6). Both regions are located within 15 arcmin from Sgr A? , thus they have excellent statistics because of the large exposure. In green and blue are the pn spectra of G0.687-0.146 and Sgr B1 soft, respectively, both are located further out, thus having lower exposure. We first selected the standard broad energy bands for the continuum with the softer band being: E = 0.5 − 2 keV; the medium 2 − 4.5 keV and the hard 4.5 − 12 keV (see Tab. 1). We note that, at high energies, the EPIC-pn camera shows strong instrumental background emission lines due to Ni Kα (at E ∼ 7.47 keV), Cu Kα (∼ 8.04 keV) and ZnCu (∼ 8.63 and 8.87 keV) that strongly contribute to the observed X-ray emission in the hard band (see Freyberg et al. 2004). To avoid contamination from these strong internal background lines, we do not consider (for the EPIC-pn images) photons in the 7.2-9.2 keV range (see Fig. 2 and Tab. 1). We chose these broad energy bands because they are typically used as input by the standard XMM-Newton point source detection algorithm and for comparison to other similar surveys of nearby galaxies (e.g. M33: Misanovic et al. 2006; Tüllmann et al. 2011; M31: Henze et al. 2014; Stiele et al. 2011; LMC: Haberl et al. 1999; SMC: Haberl et al. 2012; Sturm et al. 2013). However we note that, given the typical GC neutral column density of several 1022 cm−2 , the low energy absorption cut-off occurs at the highest energies of the standard soft band, making standard broad band RGB images 4 G. Ponti et al. TOTAL CMZ SCAN 0.400 0.200 Galactic latitude 0.000 -0.200 -0.400 -0.600 -0.800 1.500 1.000 0.500 0.000 359.500 359.000 Galactic longitude 0.00e+00 6.48e+03 1.72e+04 3.50e+04 6.41e+04 1.13e+05 1.92e+05 3.23e+05 5.42e+05 9.00e+05 1.49e+06 NEW CMZ SCAN 0.400 0.300 Galactic latitude 0.200 0.100 0.000 -0.100 -0.200 -0.300 -0.400 1.500 1.000 0.500 0.000 359.500 359.000 Galactic longitude 0.00e+00 6.48e+03 0.400 1.72e+04 3.50e+04 6.41e+04 1.13e+05 1.92e+05 3.23e+05 5.42e+05 9.00e+05 1.49e+06 OLD CMZ SCAN 0.300 Galactic latitude 0.200 0.100 0.000 -0.100 -0.200 -0.300 -0.400 1.500 1.000 0.500 0.000 359.500 359.000 Galactic longitude 0.00e+00 6.48e+03 1.72e+04 3.50e+04 6.41e+04 1.13e+05 1.92e+05 3.23e+05 5.42e+05 9.00e+05 1.49e+06 Figure 1. (Top panel) Combined exposure map of all the XMM-Newton EPIC-pn + MOS1 + MOS2 observations within one degree from Sgr A? . Such as done in Sturm et al. (2013), EPIC-MOS1 and -MOS2 exposure is weighted by a factor of 0.4 relative to EPIC-pn to account for the lower effective area. The exposure times, thus, correspond to the equivalent total EPIC-pn exposure time. Regions with less than 7.2 ks of equivalent EPIC-pn exposure have been masked out. The cleaned EPIC-pn equivalent exposure time is reported in seconds. (Medium panel) Similar exposure map for the observations of the new CMZ scan only. (Bottom panel) Similar exposure map for the old CMZ scan only (regions with less than 7.2 ks of equivalent EPIC-pn exposure are included). The maximum exposure times are ∼ 1.5 Ms, ∼ 190 ks and ∼ 45 ks during the total, new and old scan, respectively. poorly sensitive to column density fluctuations. For this reason we define a second set of broad bands, the ”GC continuum bands” (see Tab. 1). The first band (E = 0.5 − 1.5 keV) contains mainly emission from foreground sources. The second band (E = 1.5 − 2.6 keV) is selected in order to contain the low-energy GC neutral absorption cut-off, thus making it more sensitive to either column density or soft gas temperature variations. While the ”GC medium” (E = 2.6−4.5 keV) and the ”GC hard” (E = 4.5−12 keV) bands are similar to the standard broad bands. We also selected images at the energies of the soft emission lines, such as Si XIII, S XV, Ar XVII and Ca XIX. To perform continuum subtracted line intensity maps as well as line equivalent width c 0000 RAS, MNRAS 000, 1–?? 5 0(%1(((%% !$%56%% !$%112%% 0%12%% 3#%12((%%4,%1(1%% Soft 0.5-2 Standard continuum bands: Medium Hard† 2-4.5 4.5-12 !$!# Fore 0.5-1.5 78%!$5%% GC continuum bands: GC Soft GC Medium GC Hard† 1.5-2.6 2.6-4.5 4.5-12 Soft emission lines: S XV Ar XVII Ca XIX 2.35-2.56 3.03-3.22 3.78-3.99 Continuum subtraction soft emission lines: Red-Si Si-S S-Ar Ar-Ca Blue-Ca 1.65-1.77 2.1-2.3 2.7-2.97 3.27-3.73 4.07-4.5 Si XIII 1.80-1.93 !$!% &'()*+,-./01'2&3404!!05.6!!0*(1),&!" 4";<%!$5% 9":% *$'% -,#'% !"#$% 0":%% +,#'%% &$'()*%% .$#/% -,#'%% &$'()*%-,#'%% ! " Fe Kα 6.3-6.5 # 7&.(890:5.6; CFeK Figure 2. EPIC-pn spectra of the regions: G0.11-0.11 (red), Center Superbubble (black), G0.687-0.146 (green) and Sgr B1 soft (blue). In dark green are the energy bands of the broad GC continuum. In orange (bottom right) is shown the part of the hard energy band excluded in order to avoid contamination by Ni Kα, Cu Kα and ZnCu instrumental background emission lines. Yellow stripes show the energy bands of the soft and Fe K emission lines. Blue stripes indicate the regions selected for the determination of the amount of continuum underlying the soft lines. The dotted lines, from top to bottom, show the predicted emission of a source with a power-law spectrum (with slope Γ = 1.6) if absorbed by a column density of NH = 3, 5 and 9 × 1022 cm−2 , respectively. The blue and orange labels indicate the selected broad energy bands for the determination of the continuum underlying the Fe K line emission. 5-6.1 Table 1. Energy bands used for each of the different continuum, and narrow line images. Also shown are the energy bands used to determine the continuum underlying the line emission. Units are in keV. Several energy bands, at lower energies compared to the FeK lines, have been computed to determine the best continuum subtraction for the FeK lines. †To avoid contribution from the strong internal detector background emission lines, present in the EPIC-pn camera (such as: Ni Kα, Cu Kα and ZnCu), we do not consider photons in the 7.2–9.2 keV from this instrument (on the other hand, we consider such photons detected in the EPIC-MOS cameras). 2.2.1 maps, it is essential to measure the level of the continuum underlying the emission line. Therefore, we created also several images in energy bands on each side of the soft emission lines2 (selecting, as far as possible, energy ranges free from line emission; see Fig. 2 and Tab. 1). In the Fe K region we selected two energy bands for the Fe Kα and Fe XXV emission. At energies higher than Fe XXV the presence of both Fe Kβ, Fe XXVI, and of the Fe K edge can give a significant contribution. At even higher energies (E ∼ 7.5 − 8 keV) the contribution from internal background emission line (in the EPICpn camera) becomes very important, thus we decided to determine the continuum emission underlying the Fe K line emission (important to determine the Fe K line intensities and equivalent widths) through the extrapolation of the continuum red-ward of the Fe K lines (see Fig. 2 and Tab. 1). The Fe K line emission and its variations will be the prime scientific focus of two future publication (Ponti et al. in prep.; Soldi et al. in prep.; see also Ponti et al. 2014; Soldi et al. 2014) and will not be discussed here any further. All images were exposure corrected and, to remove readout streaks, the images from EPIC-pn were corrected for out-of-time events. Noisy CCDs in the MOS data (Kuntz & Snowden 2008) have been searched with the SAS task emtaglenoise and removed from the mosaic images. The energy band red-ward of the Si XIII line extends only down to 1.65 keV, because of the presence of the strong Al Kα background emission line at E ∼ 1.49 keV (Freyberg et al. 2004). c 0000 RAS, MNRAS 000, 1–?? Stray-light rejection Because the GC region is crowded with many bright (transient) Xray sources, several observations, including the new XMM-Newton scan, are badly affected by stray-light (see Fig. 3). Stray-light is produced by photons from sources located outside of the XMMNewton EPIC instrument’s fields of view and singly reflected by the mirror hyperbolas, thus creating concentric arc-like structures in the detector plane (see XMM-Newton user handbook). The straylight contribution is small (the effective collecting area for straylight is less than ∼ 0.2 % of the effective on-axis area), but a very bright source can have an important impact up to ∼ 1.4 deg outside the field of view. Analogously to the removal of bright transients, we masked the strongest stray-light artefacts in the images of individual observations. In most cases, affected regions are covered by other unaffected observations, thus leaving no features in the final mosaic map. To remove a stray-light artefact, we defined a rough region including the artefact and an individual cut-off value of the surface brightness. Using this cut-off, we created a mask from an image of this region in the total energy band that has been smoothed with a Gaussian kernel with a FWHM of 1000 beforehand. This mask was multiplied by all images and exposure maps of this observation. 2.2.2 2 Fe K lines: Fe XXV 6.62-6.8 Continuum subtraction Fe K: CsFeK CmFeK ChFeK soft medium hard 4.0-4.7 4.7-5.4 5.4-6.1 Adaptive smoothing All images have been smoothed separately using the SAS tool AS MOOTH . To prevent different smoothing patterns from introducing colour artefacts in RGB images, we adaptively smoothed all energy bands in such images with the same smoothing template. For the broadband XMM-Newton continuum RGB images (see Fig. 3, G. Ponti et al. 1.000 1.500 -0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 TOTAL CMZ SCAN Red: 0.5-2 keV Green: 2-4.5 keV Blue: 4.5-12 keV 0.500 Galactic longitude 0.000 359.500 359.000 6 Galactic latitude Figure 3. Standard broad energy band (see Tab. 1 and Fig. 2) RGB mosaic image of all XMM-Newton observations within one degree of Sgr A? (see Tab. 8). This represents the deepest X-ray view of the CMZ region with exposure higher than 0.2 Ms along the Galactic disc and 1.5 Ms in the center (see Fig. 1). X-ray emission from X-ray binaries, star clusters, supernova remnants, bubbles and superbubbles, HII regions, PWNs, non-thermal filaments, nearby X-ray active stars, the supermassive BH Sgr A? and many other features are observed (see Fig. 5 and 6). The detector background has been subtracted and adaptive smoothing applied. Residual features and holes generated by correction of the stray light from GX 3+1 are visible (see also Fig. 1) at Galactic latitudes between c 0000 RAS, MNRAS 000, 1–?? l ∼ 1.2◦ and l ∼ 1.4◦ and latitudes b ∼ −0.2◦ and b ∼ 0.4◦ . 1.000 1.500 -0.400 -0.300 -0.200 -0.100 0.000 0.100 0.200 0.300 0.400 Red: 0.5-2 keV Green: 2-4.5 keV Blue: 4.5-12 keV 0.500 Galactic longitude 0.000 359.500 359.000 7 Galactic latitude Figure 4. Standard broad energy band RGB image of the XMM-Newton CMZ scan performed in 2012. The CMZ is observed with a uniform exposure (see Fig. 1). c 0000 RAS, MNRAS 000, 1–?? 8 G. Ponti et al. Figure 5. Finding chart. The brightest X-ray point sources (all X-ray binaries) are labelled in white (see Tab. 2). In red the positions of some star clusters are reported, which are placed either in the GC or along the spiral arms of the Galaxy (see Tab. 3). With yellow dashed lines the location of some molecular complexes are shown. See WWW. MPE . MPG . DE / HEG / GC / for a higher resolution version of this figure. 4, 5, 6, 7, 15 and 17), we required a minimum signal-to-noise ratio of 6 in the 0.5-12 keV energy band image (i.e., the sum of the three energy bands composing the RGB image), as well as the standard minimum and maximum size of the smoothing Gaussian kernel of 1000 (full width half maximum) and 20000 , respectively. The signal to noise ratio at each pixel is defined as the value at that pixel divided by its standard deviation and the adaptive smoothing that we applied is making the signal to noise ratio as close at possible to 6, therefore fainter or less exposed areas are more smoothed than brighter or better exposed regions. For the narrower-band soft-line images (see Fig. 10, 11, 19 and 20), we use the S xv map as a template, requiring a minimum signal-to-noise ratio of 4 and the same standard minimum and maximum of the smoothing kernel. The same smoothing kernel is then applied to all the other bands of the RGB images. 2.2.3 Internal particle background subtraction Unless otherwise specified, internal particle background has been removed from each broad-band image. Following Haberl et al. (2012) we first create, for each selected energy band, both the total emission and the filter wheel closed images. We then re-normalise and subtract the filter wheel closed images from the total emission images. The filter wheel closed image re-normalisation factor is computed by equating, for each instrument, the number of photons in the unexposed corners of the detector to that in the filter wheel closed images (see Haberl et al. 2012 for more details). This procedure is reliable and accurate for reasonably long exposures (t ' 5 − 10 ks). For this reason datasets with total clean EPIC-pn exposure shorter than 5 ks have not been considered in this analysis. 2.2.4 Continuum subtraction To subtract the continuum emission from an emission-line image, we define a narrow band (B) containing the line, typically sandwiched by two nearby but generally wider energy bands (A and C) that are dominated by continuum emission. Under the assumptions that the emission in the A and C bands is dominated by the continuum and that the continuum emission can be described by a simple power-law, we could in principle determine the intensity of the continuum for each pixel of the band B image. Indeed, using the fluxes in A and C bands, it is possible to derive the continuum parameters (spectral index Γ and intensity). However, this requires the solution of non-linear equations. Therefore, we prefer to implement a different technique based on interpolation. Using Xspec we simulate, for power-law spectral indices going from Γ = 0.3 to 3.6, the ratio between the observed flux (e.g., number of photons measured) in the continuum in bands A and C, compared to the simulated continuum flux in band B (e.g. NB /(2 × NA ) and NB /(2 × NC )). We record these ratios and then plot them as a function of the hardness ratios (NC − NA )/(NC + NA )), which is a proxy for the spectral index Γ. We then find the best-fitting linear relationship between these values, thus determining ConAB and LinAB that are then allowing us to measure the continuum emission underlying the line emission in band B (NB ) from the intensity in band A (NA ) and the hardness ratio (NB = 2 × NA × [ConAB + LinAB × (NC − NA )/(NC + NA )]. To reduce the uncertainties, we perform the same procedure for band C, determining c 0000 RAS, MNRAS 000, 1–?? 9 Figure 6. Finding charts. (Top panel) Broadband X-ray continuum image. White ellipses show the position and size of known, radio-detected supernova remnants. Cyan ellipses indicate the position and size of bright diffuse X-ray emission possibly associated with supernova remnants that lack a clear radio counterpart (or such in the case of G359.12-0.05 that show X-ray emission significantly displaced from the radio emission associated to the radio remnant G359.07-0.02). The magenta ellipses show the location and dimension of some bright HII regions, while the red ellipses indicate some of the largest nonthermal filaments detected in radio (see Tab. 4). Blue ellipses show some PWN and the yellow dashed ellipses show the regions used to accumulate the spectra shown in Fig. 2. (Bottom panel) 90-cm radio image of the CMZ region obtained with the VLA (courtesy of LaRosa et al. 2000). For display purposes the radio supernova remnants are shown with green ellipses. The other regions have the same colour code as the top panel. The image shows the radio flux (Jy beam−1 unit). See WWW. MPE . MPG . DE / HEG / GC / for a higher resolution version of these figures. c 0000 RAS, MNRAS 000, 1–?? 10 G. Ponti et al. Figure 7. Finding chart. Zoom of the central ∼ 10 arcmin of the Milky Way as seen by XMM-Newton (same energy bands and color scheme as in Fig. 3). The position of Sgr A? is indicated by the blue cross. The red ellipses show the position and extent of filamentary and diffuse X-ray emission features associated with, e.g., non-thermal filaments (Tab. 4). The magenta dashed ellipses show the location and extension of the 20 pc bipolar X-ray lobes. The black dashed ellipses indicate the position and orientation of the CND. See WWW. MPE . MPG . DE / HEG / GC / for a higher resolution version of this figure. ConCB and LinCB . We then average these values obtaining, for each pixel: NB = NA × [ConAB + LinAB × (NC − NA )/(NC + NA )] + NC × [ConCB + LinCB × (NC − NA )/(NC + NA )]. We finally subtract this continuum emission image from the total emission image B to determine the line intensity map. 3 THE XMM-Newton BROADBAND VIEW OF THE GALACTIC CENTRE Figure 3 shows the broad energy band mosaic image of all existing XMM-Newton observations within 1 degree from Sgr A? . Figure 4 shows the Galactic centre image obtained only with data from the 2012 XMM-Newton campaign. At the GC distance of 7.8 kpc c 0000 RAS, MNRAS 000, 1–?? 11 Figure 8. Finding chart. Chandra RGB image of all the ACIS-I observations pointed at Sgr A? (see Clavel et al. 2013 for data reduction and details on the image production). The red, green and blue images show the GC soft (1.5-2.6 keV), GC medium (2.6-4.5 keV) and GC hard (4.5-8 keV) energy bands, respectively. The same regions displayed in Fig. 7 are evidenced here. See WWW. MPE . MPG . DE / HEG / GC / for a higher resolution version of this figure. (Boehle et al. 2015), 1 arcmin corresponds to 2.3 pc, 10 pc subtends ∼ 4.3’ and ∼ 0.2◦ corresponds to 28 pc. In red, green and blue, the soft (0.5-2 keV), medium (2-4.5 keV) and hard (4.5-12 keV) continuum bands are shown, respectively. Hundreds of point sources and strong diffuse emission are clearly observed in the map. These point sources are characterised by a wide variety of colours, ranging from distinctive red to dark blue. c 0000 RAS, MNRAS 000, 1–?? 3.1 Bright and transient GC point sources during the new (2012) XMM-Newton CMZ scan Many X-ray point sources are clearly visible in Fig. 3 and 4. A detailed catalogue of the properties of all the detected point sources is beyond the scope of this paper. Here we briefly describe the brightest GC sources detected by XMM-Newton and the X-ray transients 12 G. Ponti et al. G359.969-0.027 -0.010 -0.020 G359.897-0.023 G359.959-0.027 G359.921-0.030 G359.941-0.029 -0.030 G359.969-0.033 G359.933-0.037 Galactic latitude G359.971-0.038 G359.95-0.04 G359.983-0.040 -0.040 G359.933-0.039 G359.945-0.044 G359.942-0.045 Sgr A* G359.944-0.052 G359.964-0.053 -0.050 CND G359.904-0.047 G359.925-0.051 G359.956-0.052 G359.921-0.052 G359.965-0.056 G359.915-0.061 -0.060 G359.962-0.062 G359.899-0.065 -0.070 G359.977-0.076 359.980 359.970 359.960 359.950 359.940 359.930 359.920 359.910 359.900 Galactic longitude Figure 9. Zoom of the finding chart displayed in Fig. 8. in the field of the 2012 scan (see Degenaar et al. 2012 for a compendium of previously noted transients). The brightest GC point source of the 2012 XMM-Newton CMZ scan is 1E 1743.1-2843 (Porquet et al. 2003; Del Santo et al. 2006) a persistently accreting neutron star binary detected at an observed flux level of F2−10keV ∼ 1.1 × 10−10 erg cm−2 s−1 (implying an unabsorbed flux of F2−10keV,unab ∼ 2.6 × 10−10 erg cm−2 s−1 ; NH ∼ 2 × 1023 cm−2 ; obsid: 0694641201). During the 2012 XMM-Newton campaign we also detected an outburst from a new, very faint X-ray transient that we name XMMU J174505.3291445. The source has a typical quiescent luminosity at or below LX ∼ 1033 erg s−1 , but on 2012 August 31st (during obsid 0694640201) it was observed to go into outburst for about ∼ 2 hr and to reach a peak X-ray luminosity of LX ∼ 1035 erg s−1 . The detailed spectral and multiwavelength analysis of this new transient will be presented in a separate paper (Soldi et al. in prep.; but see also Soldi et al. 2014). Another faint X-ray transient, XMM J174457-2850.3, is detected in two 2012 observations; obsid: 0694641101 - 0694640301. The observed 2-10 keV flux is 1.1 ± 0.3 and 2.9 ± 0.6 × 10−13 erg cm−2 s−1 , respectively. A power-law fit to the spectrum with the photon index fixed to the value reported by Sakano et al. (2005) yields a column density of NH = (1.4 ± 0.4) × 1023 cm−2 . This source was discovered in 2001 by Sakano et al. (2005) who re- ported a tentative detection of an X-ray pulsation of ∼ 5 s, during the ∼ 25 ks XMM-Newton observation. Both a visual inspection and timing analysis of the X-ray light curve show no evidence for bursts and/or dips. However, even considering the 4 times longer exposure of the new data, we cannot exclude or confirm the ∼ 5 s periodic modulation because of the lower flux observed. In fact, during the 2001 XMM-Newton observation (obsid: 0112972101) XMM J174457-2850.3 had a flux about 10 − 40 times higher (∼ 45 × 10−13 erg cm−2 s−1 ) than in 2012 (in quiescence XMM J174457-2850.3 has a typical 2-10 keV flux lower than 0.2×10−13 erg cm−2 s−1 ). Only upper limits are measured for the other well-known X-ray transients within the field of view. The two bursters GRS 1741.9-2853 and AX J1745.6-2901 (Sakano et al. 2002; Trap et al. 2009; Ponti et al. 2014; 2015) have flux limits of F2−10keV < 2 × 10−14 erg cm−2 s−1 and F2−10keV < 10−13 erg cm−2 s−1 (obsid: 0694641101, 0694640301), respectively. Closer to Sgr A? , we find an upper limit on the 2-10 keV flux of F2−10keV < 5 × 10−12 erg cm−2 s−1 toward three other sources: CXOGC J174540.0-290031, the low-mass X-ray binary showing X-ray eclipses (Porquet et al. 2005; Muno et al. 2005), CXOGC J174540.0-290005 (Koch et al. 2014), and the magnetar discovered on April 25, 2013 (Degenaar et al. 2013; Dwelly & Ponti 2013; Mori et al. 2013; Rea et al. 2013; Kaspi et al. 2014; c 0000 RAS, MNRAS 000, 1–?? 13 Coti-Zelati et al. 2015), located at distances from Sgr A? of only ∼ 2.9, ∼ 23 and ∼ 2.4 arcsec, respectively. Finally we observe that both XMMU J174554.4-285456, the faint transient with a possible pulsation period of about 172 s (Porquet et al. 2005), and SAX J1747.7-2853, the bursting (showing also superbursts) X-ray transient (Wijnands et al. 2002; Natalucci et al. 2004; Werner et al. 2004), have flux limits of F2−10keV < 2 × 10−13 erg cm−2 s−1 . Bright sources outside the 2012 scan Three very bright sources are outside the field of view during the 2012 CMZ scan, however they imprint their presence through bright stray-light arcs. The arc features between and south of the Sgr A and C complexes (l ∼ 359.6 − 359.9◦ , b ∼ −0.15 − 0.4◦ ) testify that the bright X-ray burster 1A 1742-294 (Belanger et al. 2006; Kuulkers et al. 2007) was active during the 2012 XMMNewton campaign. The very bright arcs east of the Sgr D complex (obsid: 0694641601) are most probably produced by the very bright neutron star low-mass X-ray binary GX 3+1 (Piraino et al. 2012) located about 1.18◦ northeast of the arcs3 . On the far west edge of the 2012 scan a brightening is observed. This is due to 1E 1740.7-2942 (Castro et al. 2013; Reynolds & Miller 2010; Natalucci et al. 2014), a bright and persistent accreting microquasar, at only ∼ 1.5 arcmin from the edge of the 2012 field of view (see Fig. 3). The lack of straylight south of the Sgr B region suggests that the BH candidate IGR J17497-2821 (Soldi et al. 2006; Paizis et al. 2009) was in quiescence during these observations. Two bright X-ray bursters have been active during the 2003 XMM-Newton observation pointed to the pulsar wind nebula called The Mouse, i.e., SLX 1744-299 and SLX 1744-300 (Mori et al. 2005). 3.2 Very soft emission: Foreground emission Despite the presence of distinctively soft (red) point sources, Fig. 3 shows no strong, diffuse, very soft X-ray emission. This is mainly due to the very high column density of neutral hydrogen toward the GC (with typical values in the range NH ∼ 3 − 9 × 1022 cm−2 ; see also §7). Almost no Galactic centre X-ray radiation reaches us below E < ∼1.3, 1.7 or 2.3 keV for column density values of NH ' 3, 5, or 9 × 1022 cm−2 , respectively (see Fig. 2). The majority of the ”red” sources present in the 0.5-1.5 keV band are point-like and are associated with foreground active stars characterised by an unabsorbed soft X-ray spectrum. Two clearly extended and soft X-ray emitting sources are present in Fig. 3. These correspond to Sh2-10 and Sh2-17 (Wang et al. 2002; Dutra et al. 2003; Law et al. 2004; Fukuoka et al. 2009), two stellar clusters located in one of the Milky Way spiral arms and thus characterised by a lower column density of absorbing material, consequently appearing stronger in the 1-2.5 keV range (visible in Fig. 3 with orange colours). 3.3 Soft and hard GC emission Galactic centre radiation with energies above ∼ 2 − 3 keV can typically reach us and be detected (in green and blue in Fig. 3). GC sources with a significant continuum component (e.g. power-law or 3 This region is covered only by the observations of the 2012 XMM-Newton scan, therefore the removal of the stray-light arcs generates regions with null exposures in the final mosaic maps (e.g., Fig. 3 and 4). c 0000 RAS, MNRAS 000, 1–?? Bremstrahlung), such as observed from most GC point sources, the GC stellar clusters (e.g. the Arches, the Quintuplet and the Central cluster) as well as some supernova remnants (such as SNR G0.9+0.1 and Sgr A East) appear with a bright light blue colour. Colour gradients confirm the presence of at least two components of the diffuse emission, each having a different spatial distribution (see Figs. 3 and 4). One component dominates the emission in the soft and medium energy bands, thus appearing with a distinctively green colour. Its distribution appears to be very patchy, peaking typically at the position of known supernovae remnants. Another, harder component appears with a dark blue colour (Fig. 3). This harder emission is known to consist of at least two separate contributions. One, which is associated with intense high-ionisation Fe K lines, is smoothly distributed and peaks right at the GC; it is likely produced by faint point sources (Muno et al. 2004; Revnivtsev et al. 2009; Heard & Warwick 2013a). The other, which is associated with neutral Fe K emission lines, has a patchy distribution peaking at the position of molecular cloud complexes; it is likely due to an ensemble of X-ray reflection nebulae (see Ponti et al. 2013 for a review). We also note that the Galactic plane emission is dominated by dark blue colours (Fig. 3), while regions located ◦ ◦ at b> ∼0.2 and b< ∼ − 0.35 have a significantly greener colour. We address this in more detail in § 5, 7 and 8.7. 4 SOFT LINE EMISSION Figure 2 shows the spectra of the diffuse emission from the regions marked in magenta in Fig. 6. The ∼ 1.5 to ∼ 5 keV band shows strong, narrow emission lines, the strongest of which are Si XIII, S XV , Ar XVII and Ca XIX . This line emission, as well as the underlying continuum and the intra-line emission, are typically well fitted by a thermal model (e.g. APEC in X SPEC) with temperatures in the range 0.6 − 1.5 keV (Kaneda et al. 1997; Tanaka et al. 2000; Muno et al. 2004; Nobukawa et al. 2010; Heard & Warwick 2013b). At higher energies a power-law component with intense Fe XXV and Fe XXVI lines is also observed over the entire GC region. Additionally, neutral Fe Kα and Kβ lines are also observed. The neutral Fe K emission lines are associated with different processes, therefore they will be the focus of separate publications. 4.1 RGB images of soft emission lines The top and bottom panels of Fig. 10 show the line (continuum non-subtracted) RGB image and the inter-line continuum RGB image (see caption of Fig. 10 and Tab. 1 for more details). We note that the soft X-ray line image shows very strong colour gradients (less dramatic colour variations are observed in the continuum image). In particular, the sources DS1 (the core of Sgr D), the western part of Sgr B1 (i.e., G0.52-0.046, G0.570-0.001), Sgr C, as well as the Chimney above it, all have a distinctively green-blue colour, while G359.12-0.05, G359.10-0.5, G359.79-0.26, G359.730.35 and the entire G359.77-0.09 superbubble are characterised by orange-brown colours. G0.1-0.1, the Radio Arc, the arched filaments (see Fig. of Lang et al. 2002), G0.224-0.032, and G0.40-0.02 are also characterised by red-brown colours, however here a gradation of white and green is also present (please refer to Tab. 3 and 4 and Figs. 5 and 6 for the positions of the regions listed here). The lobes of Sgr A appear with a whiter colour than the surroundings. In addition, we observe bright red-brown emission along two broad, linear ridges having relatively sharp edges to the northwest 14 G. Ponti et al. Red: Si xiii Green: S xv Blue: Ar xvii 0.400 Galactic latitude 0.200 0.000 -0.200 -0.400 -0.600 1.500 1.000 0.500 0.000 359.500 359.000 0.000 359.500 359.000 Galactic longitude Red: Si-S Green: S-Ar Blue: Ar-Ca 0.400 Galactic latitude 0.200 0.000 -0.200 -0.400 -0.600 1.500 1.000 0.500 Galactic longitude Figure 10. (Top panel) Soft lines (continuum unsubtracted) RGB image of the CMZ. The Si XIII line emission is shown in red, S XV in green and Ar XVII+ Ca XIX in blue. (Bottom panel) RGB image of the energy bands between soft emission lines. The Si-S band emission is shown in red (see Tab. 1 for a definition of the energy bands), S-Ar in green and Ar-Ca + Blue-Ca in blue. The diffuse emission in this inter-line continuum is very similar to the soft emission line one, suggesting that the same process is producing both the lines and the majority of the soft X-ray continuum. The colour variations within the map are modulated primarily by abundance variations (top), temperature of the emitting plasma, continuum shape and absorption. and northeast of Sgr A? . This latter feature is discussed in detail in section 8.4. In spite of the fact that these images show different components (one being dominated by emission lines, the other by continuum emission), they are remarkably similar. No clear diffuse emission component is present in one and absent from the other image. This indicates that most of the diffuse soft X-ray continuum and line emission are, indeed, produced by the same process. In addition, the differences in the ratio between photons emitted in the lines and in the continuum can add valuable information for understanding the radiative mechanism. In fact, such differences could be due, for example, to different cosmic abundances and/or variations in the relative contributions of various thermal and nonthermal radiation mechanisms. In order to better highlight these differences, we map the sum of the interline continua in the same image (see caption of Fig. 11 and Tab. 1). As expected, none of the point sources is a strong soft line emitter (they in fact appear brighter in the interline image). We also note that the intense soft X-ray emitting regions in the Galactic plane, such as Sgr D, Sgr B1, Sgr C, the Chimney and G359.9-0.125 are characterised by distinctively orange-red colours, indicating they are strong line emitters. For an alternative perspective, the bottom panel of Fig. 11 shows the Si XIII + Si-S bands in red, S-Ar + S XV + Ar-Ca in green, and CFeK in blue. These energy bands are chosen to highc 0000 RAS, MNRAS 000, 1–?? 15 Red: Si xiii + S xv + Ar xvii + Ca xiv Green: Si-S + S-Ar + Ar-Ca + B-Ca Blue: 4.5-12 keV 0.400 Galactic latitude 0.200 0.000 -0.200 -0.400 -0.600 1.500 1.000 0.500 0.000 359.500 359.000 0.000 359.500 359.000 Galactic longitude Red: Si xiii + Si-S Green: S-Ar + Ar xvii + Ar-Ca Blue: 5-6.1 keV 0.400 Galactic latitude 0.200 0.000 -0.200 -0.400 -0.600 1.500 1.000 0.500 Galactic longitude Figure 11. (Top panel) RGB image composed of summed, continuum-subtracted line emission (Si XIII + S XV + Ar XVII + Ca XIX) in red, the sum of the interline continua (Si-S + S-Ar + Ar-Ca + Blue-Ca) in green and the CFeK emission in blue (see Tab. 1). (Bottom panel) RGB image, in red the Si XIII + Si-S emission, in green the S-Ar + S XV + Ar-Ca and in blue the CFeK emission. light any energy dependence in the soft X-ray emission that could be due to column density variations of the obscuring matter or temperature fluctuations of the emitting gas. In fact, the softer energy bands (Si XIII + Si-S) will be more affected by absorption or low temperature plasma emission compared to the medium (S-Ar + S XV + Ar-Ca) or high energy bands. We defer the detailed discussion of the features present in these images to the discussion of the various physical components presented in § 8 and subsections. 4.2 Continuum subtracted soft emission line maps and profiles From top to bottom, the panels of Fig. 12 show the continuumsubtracted Si XIII, S XV, Ar XVII, Ca XIX intensity maps. Although c 0000 RAS, MNRAS 000, 1–?? the continuum subtraction procedure should naturally remove the emission from the line-free point sources, small fluctuations in the continuum subtraction, in the case of the brightest sources, sometimes leave significant residuals. For this reason, we have masked out the brightest point sources in our computation of these maps. The different curves of Fig. 13 show the continuum-subtracted emission profiles (integrated over latitude from the magenta rectangle in Fig. 12) for the individual soft emission lines. The same four line profiles are compared in Fig. 13 with similar profiles in which the contribution of specific bright structures has been removed. 16 G. Ponti et al. 0.000 Si xiii 1.000 0.500 0.000 359.500 S xv Ar xvii Ca xix Figure 12. From top to bottom, continuum subtracted Si XIII, S XV, Ar XVII, Ca XIX intensity maps of all the stacked XMM-Newton observations of the CMZ. 5 SPECTRAL DECOMPOSITION In order to better trace the relative contributions of the diffuse thermal (soft and hot) and non-thermal components, we have performed a simple component separation using a list of images depicting various energy bands. We use a total of 17 energy bands: 11 for the continuum4 and 6 for the lines (tracing Si XIII, S XV, Ar XVII, Ca XIX, Fe Kα and Fe XXV, see Tab. 1). This treatment of the data allows us to be more confident about the spectral decomposition, e.g. compared to single RGB maps, retaining most of the morphological information on sufficiently large scales (i.e. beyond few arcmin scales). The general assumption is that the emission at any position 4 The eleven continuum energy bands used are: 1.0–1.5 keV; 1.5–1.8 keV; 2.0–2.35 keV; 2.55–3.05 keV; 3.25–3.75 keV; 3.95–4.70 keV; 4.70–5.40 keV; 5.40–6.30 keV; 6.50–6.60 keV; 6.80–7.80 keV and 8.20–9.50 keV. can be represented by the linear sum of three main components, namely i) a soft plasma with a temperature of 1 keV (Kaneda et al. 1997; Bamba et al. 2002); ii) a hot plasma of temperature 6.5 keV; and iii) a non-thermal component modeled by an absorbed powerlaw plus a neutral, narrow iron line (with 1 keV equivalent width), that are subject to an additional absorbing column of NH = 1023 cm−2 . All three components are also absorbed by gas in front of the GC region and both thermal plasmas are modelled using the APE model in XS PEC. The resulting model is therefore PHABS ( APEC + APEC + PHABS ( POWERLAW + G AUSS ))) and has only three free parameters: the relative normalizations of the three components. The hot plasma component represents the emission associated with faint unresolved point sources, whose cumulative spectrum is well described by a thermal spectrum (Revnivtsev et al. 2009) plus a possibly truly diffuse hot plasma component (Koyama et al. 2007). The spectral index of the non-thermal component is assumed to be Γ = 2, consistent with the values measured through the combined c 0000 RAS, MNRAS 000, 1–?? 17 Figure 13. Longitudinal intensity profiles of the Si XIII (red), S XV (green), Ar XVII (blue) and Ca XIX (violet) emission lines, integrated over Galactic latitude within the magenta rectangular region shown in Fig. 12. spectral fits of XMM-Newton spectra with higher energy data (e.g., INTEGRAL and/or NuSTAR; Terrier et al. 2010; Mori et al. 2015; Zhang et al. 2015). The strongest assumptions in this approach are that the emission can be represented everywhere with these three components. This obviously fails on bright point sources or on regions where the emission is much hotter (e.g. Sgr East). For the soft components the even stronger assumption is that absorption to the GC is assumed to be uniform over the CMZ at a value of NH = 6 × 1022 cm?2 (Sakano et al. 2002; Ryu et al. 2009), putting aside absorption in the GC region itself. Clear column density modulations are observed towards different lines of sight (e.g. Ryu et al. 2009; Ryu et al. 2013). We tested significantly different column densities (up to NH = 1.5 × 1023 cm−2 characteristic of several GC sources; see e.g. Baganoff et al. 2003; Rea et al. 2013; Ponti et al. 2015). We found that if the soft plasma normalization is significantly modified, the overall morphology is consistent. We tested various values of the other parameters (spectral index or temperatures) and did not find strong effects on the soft plasma morphology or normalization. We first produced counts, exposure and background maps for each observation and each instrument. Background was obtained from cal-closed datasets distributed in the ESAS5 calibration database. For each observation and instrument, an average RMF is computed as well as the un-vignetted ARF. For each instrument, individual observation images were reprojected using the final image astrometry and then combined to compose a mosaic. Average 6 ARF and RMF for each instrument were obtained with the FTOOLS routines ADDRMF and ADDARF. For each pixel of the final maps, we fit the measured numbers of counts in all the energy bands and instruments with a model consisting of the three aforementioned components as well as the background events number and the Out-of-Time (OoT) events for the EPIC-pn camera. The free parameters are the normalization of each individual component. We apply Cash statistics (Cash 1979) to take into account the low statistics in each pixel. This analysis 5 6 http://xmm2.esac.esa.int/external/xmm sw cal/background/epic esas.shtml https://heasarc.gsfc.nasa.gov/ftools/ftools menu.html c 0000 RAS, MNRAS 000, 1–?? NH (1022 ) F2−4.5 /F 0 F4.5−10 /F 0 0.01 1 3 5 7 10 15 30 50 70 100 150 1 0.776 0.488 0.322 0.223 0.137 0.0698 0.0156 0.0033 0.0008 1.2 × 10−4 6.2 × 10−5 1 0.968 0.912 0.857 0.809 0.737 0.636 0.414 0.244 0.149 0.0754 0.0273 Table 5. Expected ratio of the obscured flux to the un-obscured flux for different values of the column density of obscuring material. A thermally emitting gas with temperature of kT = 1 keV (PHABS * APEC model) is assumed in the computations. The predicted observed flux in both the 2 − 4.5 and 4.5 − 10 keV bands (F2−4.5 and F4.5−10 ) is computed and compared to the respective un-absorbed (F 0) flux. allows us to perform a rough spectral decomposition, better separating the spectral emission components, although retaining the maximum spatial resolution. Figure 14 presents the map of the normalisation (in units of 10−4 times the APEC normalisation) of the soft thermal emission component. The normalisation of the soft thermal component has a distribution similar to the one traced by the soft lines and the continuum (Fig. 3, 10, 11 and 12). Enhanced high-latitude soft plasma emission is observed. The white dashed lines show the position of two sharp edges in the distribution of this high latitude emission (see also Fig. 3, 10 and 11). The white solid line shows the edge of the region having more than 7.2 ks of exposure (see Fig. 1). 6 AN ATLAS OF DIFFUSE X-RAY EMITTING FEATURES The patchy and non-uniform distribution of the diffuse emission makes the recognition of the shape, the border and connection of the different structures and components difficult. Occasionally, different works report the same X-ray feature with different names and shapes and, in extreme cases, the same X-ray emitting feature is associated with different larger scale complexes. In Tab. 3 and 4 we report all the new X-ray features discussed in this paper, plus many GC features presented in previous works. The main purpose of these tables is to provide a first step towards the building of an atlas of diffuse X-ray emitting GC features. The table is available online at: WWW. MPE . MPG . DE /HEG/GC/ATLAS GC DIFFUSE X- RAY and will be updated, should the authors be notified of missing extended features. This exercise is clearly prone to incompleteness and deficiencies, however we believe this might help in providing a clearer and more systematic picture of the diffuse X-ray emission from the GC region. The spatial location and size of all these features is shown in the finding charts in Fig. 5 and 6. 18 G. Ponti et al. 0.500 0.400 Galactic latitude 0.300 0.200 0.100 0.000 -0.100 -0.200 -0.300 -0.400 1.000 0.500 0.000 359.500 359.000 Galactic longitude 0.14 0.145 0.154 0.174 0.212 0.289 0.442 0.747 1.36 2.58 5 Figure 14. Map of the normalisation of the soft thermal gas component (in units of 10−4 times the APEC normalisation). The white lines indicate the extent of the survey having more than 7.2 ks exposure. The white dashed lines show the position of the two sharp edges in the distribution of the high latitude plasma. Some bright point sources (i.e., 1E 1743.1-2843, AX J1745.6-2901, 1E 1740.7-2942, GRS 1741.9-2853) have been removed, thereby producing artificial holes in the maps at their respective locations. 7 THE FOREGROUND COLUMN DENSITY Given the high column densities of neutral or weakly ionized material absorbing the soft X-ray radiation, it is important to estimate the effects of X-ray obscuration. For example, a molecular complex having a column density of NH ∼ 1025 cm−2 , such as the Sgr B2 core, would completely obscure the radiation below about 4 keV, if placed in front of the GC; see Fig. 2. To calculate the effects of absorption of the X-ray emission, we computed the flux generated by a thermally emitting plasma with temperature of kT = 1 keV (using a PHABS * APEC model), in both the 2 − 4.5 and 4.5 − 10 keV bands, after being absorbed by a given column density of neutral material (see also Fig. 2). For each column density explored, we report in Tab. 5 the ratio of the observed flux (F2−4.5 and F4.5−10 ) over the respective un-absorbed (F 0) flux. We note that the hard X-ray band starts to be affected (corresponding to flux reductions up to a factor of 2) for column densities up to NH ∼ 3 × 1023 cm−2 while it is heavily affected (flux reduction of a factor of 10 or more) for NH ∼ 1024 cm−2 or higher (see Tab. 5). At lower energies, the obscuration effect is even more pronounced. Already, for NH ∼ 3 × 1022 cm−2 , the 23 −2 observed flux is less than half and for NH > ∼ 5 × 10 cm it is less than 0.1 % of its un-obscured flux. This indicates that the softer band is expected to be heavily affected by absorption. density estimated from the dust has large uncertainties that can be mainly ascribed to the uncertainty associated with the dust-to-NH ratio. In particular, the column densities shown in this map appear to be systematically larger than what is measured with other methods. For example, the column density of the G0.11-0.11 massive cloud is estimated to be NH ∼ 5 − 6 × 1023 cm−2 in this map, while Amo-Baladron et al. (2009) measure NH ∼ 2 × 1022 cm−2 , through a detailed modelling of the molecular line emission. The core of Sgr B2 is estimated by Molinari et al. (2011) to have NH ∼ 3 × 1025 cm−2 , while modelling of the X-ray emission (Terrier et al. 2010) suggests NH ∼ 7 × 1023 cm−2 , more than an order of magnitude lower. Moreover, the average column densities of G0.40-0.02, G0.52-0.046, G0.57-0.018 and a fourth region (the magenta ellipse in Fig. 17) are estimated to be NH ∼ 4 × 1023 , 4 × 1023 , 1.2 × 1024 and 1.5 × 1024 cm−2 , respectively, from the dust map, while they are measured to be in the range NH ∼ 7 − 10 × 1022 cm−2 , from modelling of the X-ray emission. Therefore, the total normalisation of the NH map built from the dust emission appears to be overestimated. However, the method employed to produce it does not suffer from self-absorption, so it is presumably giving unbiased relative NH ratios. 7.2 7.1 Column density distribution The top panel of Fig. 15 shows the neutral Hydrogen column density distribution as derived from dust emission (Molinari et al. 2011)7 . The image shows the NH distribution in logarithmic scale in the range NH = 4.5×1022 −3.8×1025 cm−2 . This total column 7 We do not show the entire CMZ, because of the limited coverage of the Herschel dust emission map (Molinari et al. 2011). X-ray emission modulated by absorption The bottom panel of Fig. 15 shows the X-ray map with the columndensity contours overlaid for comparison. We observe that, as expected, no soft X-ray emission is observed toward the central part of the most massive molecular cores. In particular we observe depressed X-ray emission from: i) the Sgr B2 nucleus and its envelope (with NH > 7 × 1023 cm−2 ); ii) the almost perfect coincidence between the hole in soft X-ray emission east of G0.2240.032 (see Fig. 5 and 17) and the shape of the so-called ”Brick” molecular cloud, M0.25+0.01 (see Fig. 5; Clark et al. 2013); iii) c 0000 RAS, MNRAS 000, 1–?? 19 0.400 0.300 Galactic latitude 0.200 0.100 0.000 -0.100 -0.200 -0.300 -0.400 -0.500 0.800 0.600 0.400 0.200 0.000 359.800 359.600 359.400 Galactic longitude Figure 15. Top panel: Neutral Hydrogen column density distribution as derived from dust emission (Molinari et al. 2011). The image shows the NH distribution in logarithmic scale from NH = 4.5 × 1022 up to 3.8 × 1025 cm−2 . The green, magenta and white contour levels correspond to NH = 1.5 × 1023 , 7 × 1023 and 1.5 × 1024 cm−2 , respectively. Bottom panel: X-ray continuum RGB map (Fig. 3) with the column-density contours overlaid. c 0000 RAS, MNRAS 000, 1–?? 20 G. Ponti et al. BRIGHT AND TRANSIENT POINT SOURCES Source name Coordinates‡ Flux† within the 2012 CMZ scan 1E 1743.1-2843 XMMU J174505.3-291445 XMMU J174457-2850.3 GRS 1741.9-2853 AX J1745.6-2901 CXOGC J174540.0-290031 SGR J1745-2900 XMMU J174554.4-285456 SAX J1747.7-2853 CXOCG J174540.0-290005 0.2608,-0.0287 359.6756,-0.0634 0.0076,-0.1743 359.9528,+0.1202 359.9203,-0.0420 359.9435,-0.0465 359.9441,-0.0468 0.0506,-0.0429 0.2073,-0.2385 359.9497,-0.04269 within the total GC scan 1E 1740.7-2942 1A 1742-294 IGR J17497-2821 GRO J1744-28 XMMU J174654.1-291542 XMMU J174554.4-285456 SLX 1744-299 SLX 1744-300 359.1160,-0.1057 359.5590,-0.3882 0.9532,-0.4528 0.0445,+0.3015 359.8675,-0.4086 359.1268,-0.3143 359.2961,-0.8892 359.2565,-0.9111 110 14 0.3 < 0.02 < 0.1 <5 <5 < 0.2 < 0.2 <5 References 90,92,93 94 90,95 90,59,97 59,90,91,124 98,99 101,102,103,104 98 90,105,106,107 100 111,112,113 90,108,109 90,114,115 90 90 84,85,86 37,59,87,88,89 37,59,87,88,89 Table 2. List of bright and transient point sources during the 2012 XMM-Newton scan as well as bright point sources observed in all scans of the region (see Fig. 5). To avoid exessive crowding around Sgr A? , CXOCG J174540.0-290005 and SGR J1745-2900 are not shown. †Fluxes are given in units of 10−12 erg cm−2 s−1 and correspond to the mean flux observed during the 2012 XMM-Newton scan of the CMZ. ‡Coordinates are in Galactic format. References: (1) Wang et al. 2006a; (2) Yusef-Zadeh et al. 2002; (3) Capelli et al. 2011; (4) Tatischeff et al. 2012; (5) Sakano et al. 2003; (6) Habibi et al. 2013; (7) Habibi et al. 2014; (8) Krivonos et al. 2014; (9) Clavel et al. 2014; (10) Dutra et al. 2003; (11) Law et al. 2004; (12) Fukuoka et al. 2009; (13) Wang et al. 2002a; (14) Tsuru et al. 2009; (15) Mori et al. 2008; (16) Mori et al. 2009; (17) Heard & Warwick 2013a; (18) Maeda et al. 2002; (19) Park et al. 2005; (20) Koyama et al. 2007a; (21) Kassim & Frail 1996; (22) Nobukawa et al. 2008; (23) Senda et al. 2002; (24) Renaud et al. 2006; (25) Mereghetti et al. 1998; (26) Gaensler et al. 2001; (27) Porquet et al. 2003a; (28) Aharonian et al. 2005; (29) Dubner et al. 2008; (30) Nobukawa et al. 2009; (31) Sawada et al. 2009; (32) Morris et al. 2003; (33) Morris et al. 2004; (34) Markoff et al. 2010; (35) Zhang et al. 2014; (36) Nynka et al. 2013; (37) Gaensler et al. 2004; (38) Pedlar et al. 1989; (39) Cotera et al. 1996; (40) Figer et al. 1999; (41) Johnson et al. 2009; (42) Lu et al. 2008; (43) Lu et al. 2003; (44) Yusef-Zadeh et al. 2005; (45) Baganoff et al. 2003; (46) Ho et al. 1985; (47) Bamba et al. 2002; (48) LaRosa et al. 2000; (49) Morris & Yusef-Zadeh 1985; (50) Lang et al. 1999; (51) Anantharamaiah et al. 1991; (52) Yusef-Zadeh & Morris 1987a; (53) Yusef-Zadeh & Morris 1987b; (54) Yusef-Zadeh & Morris 1987c; (55) Muno et al. 2008; (56) Uchida et al. 1992; (57) Predehl & Kulkarni 1995; (58) Senda et al. 2003; (59) Sakano et al. 2002; (60) Coil et al. 2000; (61) Murakami 2002; (62) Yusef-Zadeh et al. 2007; (63) Dutra & Bica 2000; (64) Zoglauer et al. 2014; (65) Koyama et al. 2007b; (66) Nakashima et al. 2010; (67) Downes & Maxwell 1966; (68) Tanaka et al. 2009; (69) Tanaka et al. 2007; (70) Wang et al. 2006b; (71) Wang et al. 2002b; (72) Phillips & Marquez-Lugo 2010; (73) Hewitt et al. 2008; (74) Reich & Fuerst 1984; (75) Gray 1994; (76) Roy & Bhatnagar 2006; (77) Marquez-Lugo & Phillips 2010; (78) Borkowski et al. 2013; (79) Yamauchi et al. 2014; (80) Inui et al. 2009; (81) Green 2014; (82) Yusef-Zadeh et al. 2004; (83) Nord et al. 2004; (84) Uchiyama et al. 2011; (85) Heinke et al. 2009; (86) Muno et al. 2006; (87) Mori et al. 2005; (88) Skinner et al. 1990; (89) Pavlinski et al. 1994; (90) Degenaar et al. 2012; (91) Ponti et al. 2014; (92) Porquet et al. 2003b; (93) Del Santo et al. 2006; (94) Soldi et al. 2014; (95) Sakano et al. 2005; (97) Trap et al. 2009; (98) Porquet et al. 2005a; (99) Muno et al. 2005b; (100) Kock et al. 2014; (101) Degenaar et al. 2013; (102) Dwelly & Ponti 2013; (103) Rea et al. 2013; (104) Kaspi et al. 2014; (105) Wijnands et al. 2002; (106) Natalucci et al. 2004; (107) Werner et al. 2004; (108) Belanger et al. 2006; (109) Kuulkers et al. 2007; (110) Piraino et al. 2012; (111) Castro et al. 2013; (112) Reynolds & Miller 2010; (113) Natalucci et al. 2014; (114) Soldi et al. 2006; (115) Paizis et al. 2009 (116) Lu et al. 2013; (117) Do et al. 2013; (118) Yelda et al. 2014; (119) Bamba et al. 2000; (120) Bamba et al. 2009; (121) Ohnishi et al. 2011; (122) Zhao et al. 2013; (123) Hales et al. 2009; (124) Ponti et al. 2015. the core of the Sgr C complex8 ; iv) the regions around DB-58 and at Galactic position l ∼ 0.2, b ∼ −0.48◦ also appear to have darker colours and, once again, it is possible to find molecular complexes (M0.018+0.126 and M0.20-0.48) covering roughly the same region (see Fig. 5 and 15). All these clouds are characterised by very high 23 −2 column densities NH > ∼3 − 7 × 10 cm and they most probably lie in front of Sgr A? and of most of the GC (e.g., according to the twisted ring model of Molinari et al. 2011). Therefore, they are absorbing the GC’s extended, soft X-ray emission. All this evidence suggests that at least the most massive clouds 8 At this location a sharp transition in the soft X-ray emission, with an arc-like shape, is observed. This is spatially coincident to the edge of a very dense core of dust, suggesting that the modulation in the soft X-ray emission is induced by obscuration by the molecular cloud. located in front of the GC do actually modulate (obscure) the soft X-ray emission. However, fluctuations in column densities cannot be the only cause for the observed distribution of soft X-ray emission for two reasons. First, we do observe only weak soft X-ray emission along some lines of sight having a low column density of molecular material (such as around the Sgr C and Sgr D complexes and south of the Sgr B1 region). Second, we do detect intense (among the brightest) soft X-ray emission from several regions such as the Sgr A complex and the cores of the Sgr C and Sgr D complexes, where some of the highest column-density clouds are found. In particular, in Sgr A very intense soft X-ray emission is observed along the line of sight toward the 50 km s−1 , the Bridge (Ponti et al. 2010) and the G0.11-0.11 clouds, some of the highest column density clouds in the CMZ. Although this might be explained by placing these clouds on the far side of the CMZ, it c 0000 RAS, MNRAS 000, 1–?? 21 ATLAS OF DIFFUSE X-RAY EMITTING FEATURES Name STAR CLUSTERS: Central cluster Quintuplet Arches Sh2-10[† Sh2-17[† DB-05[† Other name or associated features Coordinates (l, b) Size arcmin References G0.12+0.02 DB-6 DB-58 G0.33-0.18 359.9442, -0.046 0.1604, -0.0591 0.1217, 0.0188 0.3072,-0.2000 0.0013, 0.1588 0.31 -0.19 0.33 0.5 0.7 1.92 1.65 0.4 45,116,117,118 1,63,11 1,2,3,4,5,6,7,8,9,39,40,11 10,11,12,63,11 13,63,11 22,63,11 359.03,-0.96 359.07,-0.02 359.12,-0.05 359.10,-0.51 359.41,-0.12 359.46,+0.04 359.73,-0.35 359.77,-0.09 359.84,-0.14 359.79,-0.26 0.00,-0.16 359.94, -0.04 359.93,-0.09 359.963, -0.053 0.109,-0.108 0.13,-0.12 0.224,-0.032 0.34,+0.045 26 × 20 22 × 10 24 × 16 22 × 22 3.5 × 5.0 6.8 × 2.3 4 20 × 16 15 × 3 8 × 5.2 5.88 1 3.2 × 2.5 13.6 × 11 3×3 2.3 × 4.6 14 × 8.8 X-R 48,51,75,76,81,119,120 R 14,48,51,66 X 66 X-R 37,48,51,56,74,75,81,120,121 X 14 X 14 X 58 X 15,16,17,58 X 15,16,17,58 X 15,16,17,58 X This work X-R 32,33,34,17 R 35,38,43,47,58,60,61 X-R 5,18,19,20,48,75,81 X This work X 17 X This work R 21,48,51,81,82 0.40,-0.02 4.7 × 7.4 X 22 0.519,-0.046 0.57,-0.001 0.570,-0.018 0.61,+0.01 0.867,+0.073 1.17,+0.00 1.02,-0.17 2.4 × 5.1 1.5 × 2.9 0.2 2.2 × 4.8 7.6 × 7.2 3.4 × 6.9 10 × 8.0 This work This work X 23,24,58,59,68,80 X 22,65,79 R 25,26,27,28,29,48,75,81,82 X 31 R 30,31,48,51,75,77,81,82 10 × 10 R 73,81,82 SNR and Super-bubbles candidates: G359.0-0.9††† G358.5-0.9 - G359.1-0.9 G359.07-0.02 G359.0-0.0 G359.12-0.05 G359.10-0.5††† G359.41-0.12 Chimney[† G359.73-0.35†‡ G359.77-0.09 Superbubble G359.9-0.125 G359.79-0.26\ G0.0-0.16†\ 20 pc lobes G359.92-0.09‡ Parachute - G359.93-0.07 Sgr A East G0.0+0.0 G0.1-0.1 Arc Bubble G0.13,-0.11[ G0.224-0.032 G0.30+0.04 G0.3+0.0 G0.34+0.05 G0.33+0.04 G0.42-0.04 Suzaku J1746.4-2835.4 G0.40-0.02 G0.52-0.046 G0.57-0.001 G0.57-0.018† CXO J174702.6-282733 G0.61+0.01† Suzaku J1747.0-2824.5 G0.9+01♥ SNR 0.9+0.1 DS1 G1.2-0.0 Sgr D SNR G1.02-0.18 G1.05-0.15 G1.05-0.1 G1.0-0.1 G1.4-0.1 1.4,-0.10 Table 3. Atlas of diffuse X-ray emitting features. The first two columns in the table indicate the name primarily used in this work to refer to the feature as well as the other names used in the previous literature. The third and fourth columns show the coordinates of each feature as well as its approximate projected size. Finally, the fifth column provides references to selected works discussing the feature. For convenience, we report in Tab. 2 all the references ordered according to the numbering used in this table. The other names column shows the different designations used in previous literature. In the case of bubbles, these features are not necessarily referring to the same structure but to features forming the bubble candidate. †Possibly due to a thermal filament. ‡The interpretation as a SNR is probably obsolete. †‡ Most probably a foreground feature. \ This feature appears to be part of the superbubble G359.77-0.09. †\ New extended X-ray feature, possibly part of the superbubble G359.77-0.09. [ This feature appears to be part of the Arc bubble. Possibly connected to G0.61+0.01. ♥ X-ray emission primarily non-thermal, therefore it appears also in the next table. †† New extended X-ray feature, possibly part of the superbubble G359.77-0.09. [† The low X-ray absorption towards these star clusters indicate that they are located in front of the GC region. ††† Because of the low X-ray absorption column density (NH ∼ 2 × 1022 cm−2 ) this is most probably a foreground source (Bamba et al. 2000; 2009). [† The Chimney is most probably either part of a large scale structure (see §8.7) or an outflow from G359.41-0.12 (Tsuru et al. 2009), therefore most probably it is not a separate SNR. appears that these regions are characterised by truly enhanced soft X-ray emission (see e.g. §8.4, 8.5 and 8.6). We defer the detailed disentangling of these effects to an elaborate spectral study of these regions. 8 DISCUSSION In the process of systematically analysing all XMM-Newton observations of the central degrees of the Galaxy, we have discovered several new extended features and have produced an atlas of known, extended soft X-ray features. Here we discuss their general properties and investigate the origin/existence of several specific features. c 0000 RAS, MNRAS 000, 1–?? 8.1 General properties To compute the total observed (absorbed) flux from the CMZ we first mask out the emission from the brightest binaries, by excluding a circle with a radius of: 10 around SAX J1747.72853; 1.50 for AX J1745.6-2901 and GRS1741.9-2853; 20 for XMMU J174445.5-295044; 2.50 for 1E 1743.1-2843; 3.50 for IGR J17497-2821 and 1E 1740.7-2942 (see white circles in Fig. 5). We then compute the total observed count rate from two boxes, one with a size of 1.5◦ × 0.35◦ (l × b) and centered on Sgr A? and one with a bigger size of 2.08◦ × 0.413◦ centered on l = 0.232◦ , b = 0.080◦ . We then measure the total count rate within these re- 22 G. Ponti et al. ATLAS OF DIFFUSE X-RAY EMITTING FEATURES Name Other name or associate features Radio and X-ray filaments and PWN candidates: Snake G359.15-0.2 G539.40-0.08 G359.43-0.14 Sgr C Thread Ripple filament G359.54+0.18 G359.55+0.16 X-ray thread Suzaku J174400-2913 Crescent G359.79+0.17 Curved filament Pelican G359.85+0.47 Cane G359.87+0.44 G359.85+0.39 Sgr A-E G359.889-0.081- wisp XMM J174540-2904.5 G359.89-0.08 G359.897-0.023 G359.899-0.065 Sgr A-F G359.90-0.06 G359.904-0.047 G359.915-0.061 G359.91-1.03 G359.921-0.030 F7 G359.921-0.052 The Mouse G359.23-0.82 G359.925-0.051 G359.933-0.037 F2 G359.933-0.039 F1 G359.941-0.029 G359.942-0.045 G359.944-0.052 G359.945-0.044 G359.95-0.04 G359.956-0.052 G359.959-0.027 F5 Southern thread G359.96+0.09 359.96+0.09 G359.962-0.062 G359.964-0.053 F3 G359.965-0.056 F4 G359.969-0.033 G359.970-0.009 F8 G359.971-0.038 F6 G359.974-0.000 F9 G359.977-0.076 Cannonball J174545.5-285829 G359.983-0.040 G359.98-0.11 G0.007-0.014 G0.008-0.015 G0.014-0.054 G0.017-0.044 MC2 G0.02+0.04 G0.021-0.051 G0.029-0.08 G0.032-0.056 G0.029-0.06 - F10 G0.03-0.06 G0.039-0.077 G0.062+0.010 G0.06+0.06 Northern thread G0.09+0.17 G0.08+0.15 G0.097-0.131 Radio Arc GCRA G0.16-0.15 G0.116-0.111 G0.13-0.11 G0.15-0.07 Steep spectrum of Radio Arc XMM J0.173-0.413 G0.17-0.42 S5 G0.223-0.012 G0.57-0.018† CXO J174702.6-282733 G0.61+0.01† G0.9+0.1 Coordinates (l, b) Size arcsec References 359.15,-0.17 359.40,-0.08 359.43,-0.14 359.45,-0.01 359.548,+0.177 359.55,+0.16 312 × 54 27.5 × 5.1 21.4 × 3.9 500 × 42 320 × 55 56.1 × 8.0 R 48 X 41 X 41 R 48,51 R 43,44,48,51 X 13,41,42,43,79 359.791,+0.16 300 × 74 R 63,50,51 359.859,+0.426 359.87,+0.44 300 × 54 420 × 50 R 48,50,51 R 48 359.889,-0.081 20 × 5 R X 5,35,41,42,43,44,50,55 359.897,-0.023 359.899,-0.065 6.4 × 4 6.5 × 2.5 X 55 X 42,44,55 359.904,-0.047 359.915,-0.061 359.919,-1.033 359.921,-0.030 359.921,-0.052 359.30-0.82 359.925,-0.051 359.934,-0.0372 359.933,-0.039 359.941,-0.029 359.942,-0.045 359.944,-0.052 359.945,-0.044 359.950,-0.043 359.956,-0.052 359.959,-0.027 359.96,+0.11 6.5 × 3 7×2 138 × 36 7.5 × 3 5.5 × 2 156 × 108 8 × 2.2 12 × 3 5×2 6×2 5×3 9 × 1.5 6 × 2.5 10 × 4 4 × 2.5 9×3 500 × 40 X 55 X 55 R 48 X 42,55 X 55 PWN F 37,48,57,123 X 55 X 41,42,55 X 42,55 X 41,55 Stellar wind X 55 X 41,55 X 41,1,42,55 PWN X 55,70 PWN X 55 X 41,42,55 R 48,50,51 359.962,-0.062 359.964,-0.053 359.965,-0.056 359.969,-0.033 359.970,-0.009 359.971,-0.038 359.974,-0.000 359.977,-0.076 359.983,-0.0459 359.983,-0.040 359.979,-0.110 0.008,-0.015 0.014,-0.054 0.017,-0.044 0.0219,+0.044 0.021,-0.051 0.029,-0.08 0.0324,-0.0554 5.5 × 3.5 16 × 3.5 9×3 5×2 10 × 2.5 16 × 8 7×2 6×4 30 × 15 6.5 × 4.5 Streak 11 × 3.5 18 × 14 15 × 4 Streak 15 × 12 29 × 18 35 × 6 X 55 X 41,42,45,55 PWN X 42,55 X 55 X 41,42,55 PWN X 41,42,55 PWN X 42 X 55 X-R PWN 36,122 X 42,55 R 50 X 41,55 X 55 X 41,42 FeKa R 50 X 55 X 55 FeKa 41,42,55 PWN 0.039,-0.077 0.062,+0.010 0.09,+0.17 22 × 15 40 × 25 714 × 48 X 55 R 50,55 R 48,49,50,51 0.097,-0.131 0.167,-0.07 70 × 50 1690 × 145 X 55 R 38,48,49,50,51 0.116,-0.111 0.13,-0.11 0.138,-0.077 0.173,-0.413 0.17,-0.42 0.223,-0.012 0.57,-0.0180 0.61,+0.01 0.9,+0.1 50 × 40 55 × 12 X 55 71,17,41,42 PWN R 50 X This work 180 × 18 912 R 82 50 × 20 0.33 132 × 288 X 41 X 23,24,58,59,68,79,80 X 22,65,79 PWN Table 4. Atlas of diffuse X-ray emitting features. This table has the same structure as Tab. 3. Because of possible misplacements between the peak emission of the radio and X-ray counterparts of filaments, SNR, PWN and other diffuse structures (generally related to the different ages of the population of electrons traced at radio and X-ray bands), when available we give the best X-ray position (following the preference: Chandra, XMM-Newton, Suzaku), otherwise we state the radio position. We cite the literature results separating thes between the X-ray (X) from the radio (R) detections. †Possibly part of a young supernova remnant. For convenience, we report in Tab. 2 all the references ordered according to the numbering used in this table. c 0000 RAS, MNRAS 000, 1–?? 23 al. 1989; 2004; see figures 7a, 16b,c, and 17a,b,c of the latter reference, where the filament is labelled ”S5”), but because it is not continuous or exactly parallel with the filaments of the Arc, it is not completely evident that it is an extension of the Arc in three dimensions. The X-ray filament has a hard X-ray colour and does not appear in the soft line images, indicating a non-thermal emission spectrum. Three other nonthermal radio filaments have been found to have X-ray emission along some portion of their lengths: G359.54+0.18, G359.89-0.08, and G359.90-0.06 (Lu, Wang & Lang 2003; Sakano et al. 2003; Lu et al. 2008; Johnson et al. 2009; Morris, Zhao & Goss 2014; Zhang et al. 2014). XMM 0.173-0.413 is the only one of the four known cases where the X-ray emission is not at or near a location where the radio filament shows unusually strong curvature. 8.3 Figure 16. XMM-Newton image in the 2-12 keV band showing the new X-ray filament located south of the Radio Arc. The filament is located at l ' 0.173◦ and b ' −0.413◦ and appears as a thin (< 0.15 arcmin) and long (∼ 2.7 arcmin) filament running along the north-south direction, as the Radio Arc. The brightest source in the image is SAX J1747.7-2853. gions and convert it into a flux9 assuming that the soft X-ray diffuse emission is dominated by a thermally emitting plasma with a temperature of kT = 1 keV (Kaneda et al. 1997; Bamba et al. 2002). In particular, we assumed an APEC emission component with temperature kT = 1.08 keV, absorbed by a column density of neutral material of NH = 6 × 1022 and with Solar abundance. The fluxes of the integrated continuum and line intensities (line plus continuum) within the big and small boxes are reported in Tab. 6. This corresponds to an observed 2 − 12 keV luminosity of L2−12 = 3.4 × 1036 erg s−1 and L2−12 = 2.6 × 1036 erg s−1 for the big and small boxes, respectively, assuming a distance of 8 kpc to the Galactic center (Reid 1993; Reid et al. 2009). The top and bottom panels of Fig. 17 show the X-ray (2.5–4.5 keV) and 850 µm (Pierce-Price et al. 2000) maps of the Sgr B1 region. Four enhancements of X-ray emission are clearly present (shown by the green dashed ellipses in Fig. 17). In particular, G0.5700.001, G0.52-0.046 and G0.40-0.02 correspond to holes in the dust distribution derived from the 850 µm radiation (see bottom panel of Fig. 17). To reinforce this evidence, we observe that their X-ray edges can be traced in the dust distribution all around the X-ray enhancements, suggesting a tight connection between the two. Such phenomenology is typical of SNe exploding within or near molecular clouds and interacting with them, creating bubbles in the matter distribution (Ferreira & de Jager 2008; Lakicevic et al. 2014). Indeed, in this case, the SN ejecta might have cleared the entire region that is not filled with hot, X-ray emitting plasma, pushing away the ambient molecular material. However, this is not the only possibility. The apparent X-ray enhancements could have resulted instead from a higher obscuration surrounding the submm holes, leading to higher X-ray extinction at the edges. We note that none of these regions has a known radio SNR counterpart (see Fig. 17). However, the radio emission might be confused within the very high radio background of diffuse emission created by G0.30+0.04, the several HII regions present in this region (see Fig. 6), and the bright, extended synchrotron background of the GC. Enhanced X-ray emission has already been reported towards G0.40-0.02 (Nobukawa et al. 2008) and close to G0.570-0.001 (see cyan ellipses in Fig. 17). 8.3.1 8.2 A new X-ray filament, XMM 0.173-0.413 We observe a new X-ray filament extending ∼ 2.2 arcmin perpendicular to the Galactic plane and situated at l = 0.173◦ , b = −0.413◦ , which is almost directly toward negative longitudes from the GC Radio Arc. It coincides with the brightest segment of a much longer radio filament that extends toward the southernmost extensions of the filaments of the Radio Arc (Yusef-Zadeh et 9 To perform this task we use WEBPIMMS: https:heasarc.gsfc.nasa.govcgibinToolsw3pimmsw3pimms.pl. As explained in § 2.1, the combined count rate is the sum of the EPIC-pn plus the EPIC-MOS count rates after scaling the latter exposure maps by 0.4. We then use the EPIC-pn, medium-filter, rate-to-flux conversion computed with WEBPIMMS. c 0000 RAS, MNRAS 000, 1–?? SNR excavated bubbles within the CMZ? Spectral analysis To further investigate the origin of these structures, we extracted a spectrum from each of these features (in either obsid 0694641301 or 0694641201). We fitted each spectrum with a model composed of a SNR emission component (fitted with a PSHOCK model; a constant temperature, plane-parallel plasma shock model, meant to reproduce the X-ray emission from a supernova remnant in the Sedov phase) plus the emission components typical of the GC environment such as a hot thermal plasma (with temperature in the range: kT = 6.5 − 10 keV), and an Fe Kα emission line, all absorbed by foreground neutral material (PHABS *( PSHOCK + APEC + GAUS ) in XS PEC). We assume that all these components have Solar abundances. Possible confusion effects produce uncertainties associated with the determination of the correct sizes of these candidate SNRs, so some of the results presented here, such as the dynamical timescales, could thereby be affected. 24 G. Ponti et al. G0.30+0.04 2-4.5 keV emission 0.100 G0.224-0.032 0.050 G0.570-0.001 G0.40-0.02 Galactic latitude G0.61+0.01 0.000 G0.57-0.018 -0.050 G0.52-0.046 Tail of stars from Quintuplet -0.100 -0.150 0.600 0.550 0.500 0.450 0.400 0.350 0.300 0.250 0.200 0.150 Galactic longitude Scuba 850 micron 0.100 G0.224-0.032 Galactic latitude 0.050 G0.40-0.02 G0.570-0.001 0.000 -0.050 -0.100 G0.52-0.046 -0.150 0.600 0.550 0.500 0.450 0.400 0.350 0.300 0.250 0.200 0.150 Galactic longitude Figure 17. (Top panel) 2-4.5 keV map. The dashed white ellipse shows the location of the radio SNR G0.30+0.04. The dashed and solid cyan ellipses show the positions of known X-ray SNRs and superbubbles, respectively. The green dashed ellipses indicate enhancements of soft X-ray emitting gas (G0.40-0.02 was already observed in X-rays by Nobukawa et al. 2008), the magenta ellipse is used in the spectral analysis as a background region (Back in Tab. 7). The dashed yellow ellipse shows the location of the massive stars that could be in the tidal tail of the Quintuplet cluster (Habibi et al. 2013; 2014). (Bottom panel) 850 µm map of the GC obtained with the SCUBA bolometer (Pierce-Price et al. 2000). The four dashed ellipses indicated by the green ellipses and showing enhanced X-ray emission correspond to holes in the 2mm gas distribution. c 0000 RAS, MNRAS 000, 1–?? 25 Big box Small box Flux Surf. Bright. Flux Surf. Bright. F1−2 keV = 19.0 F2−4.5 keV = 155.0 F4.5−12 keV = 290.7 FSi xiii = 4.4 FS xv = 15.1 FAr xvii = 15.1 FCa xix = 18.4 f1−2 keV = 6.2 f2−4.5 keV = 50.7 f4.5−12 keV = 95.1 fSi xiii = 1.5 fS xv = 4.9 fAr xvii = 4.9 fCa xix = 9.9 F1−2 keV = 12.9 F2−4.5 keV = 119.3 F4.5−12 keV = 219.0 FSi xiii = 3.4 FS xv = 12.0 FAr xvii = 11.9 FCa xix = 14.3 f1−2 keV = 6.9 f2−4.5 keV = 64.3 f4.5−12 keV = 118.0 fSi xiii = 1.8 fS xv = 6.4 fAr xvii = 6.4 fCa xix = 7.7 Table 6. Fluxes (F) and surface brightnesses (f) of the continuum and the line intensities (line plus continuum) integrated over the small and big boxes described in § 8.1. The fluxes are given in units of 10−12 erg cm−2 s−1 , while the surface brightnesses are given in 10−15 erg cm−2 s−1 arcmin−2 . G0.40-0.02 shows a best-fit temperature and normalisation of the warm plasma associated with the SNR component of kT = −2 0.55 ± 0.1 keV and Apsho = 2.4+5 , while the column −1 × 10 density is observed to be NH = 7.7 ± 0.8 × 1022 cm−2 (see also Nobukawa et al. 2008). To test whether the observed X-ray enhancement is due to a real variation of the intensity of the soft X-ray emission or whether it is the product of lower extinction, we also extracted a spectrum from a nearby background comparison region (magenta in Fig. 17 and ”Back” in Tab. 7). This second region has the same size as G0.40-0.02 and it is located in a fainter region in X-rays, characterised by higher NH , as suggested by the 850 µm map (see bottom panel of Fig. 17). This background region shows a slightly higher absorption column density, NH ∼ 8.8 ± 1 × 1022 cm−2 , and no significant warm plasma component. If we impose the presence of a warm plasma component having the same temperature and τ (the ionisation timescale of the shock plasma model) as observed in G0.40-0.02 we obtain an upper limit to its normalisation of Aphsho < 6×10−3 . This suggests that the enhanced X-ray emission towards G0.40-0.02 is due to a real excess of X-ray emission and is not a simple byproduct of lower extinction. The thermal energy, the dynamical timescale and the size of G0.40-0.02 are Eth ∼ 1.9 × 1050 erg, tdy ∼ 3700 yr and 8.6 × 5.5 pc2 , respectively, as expected for a young SNR in the Sedov-Taylor phase (derived from the equations shown in Maggi et al. 2012). found neither radio nor nucleosynthetic decay products (such as 44 Ti), questioned such an interpretation. Further investigations are required to understand the link, if any, between these features. The morphology of G0.224-0.032 appears more complex, compared to the other SNR candidates. The edge of the X-ray emission is well defined only towards the Brick molecular cloud (designated M0.25+0.01 in Fig. 5) that, with its very high column density, can obscure the soft X-ray emission there. In any case, the fit of its X-ray spectrum shows parameters typical of a SNR. In particular, we derive a thermal energy and a dynamical time of Eth ∼ 2.6 × 1050 erg and tdy ∼ 1800 yr, respectively. Therefore, G0.224-0.032 might be a new SNR partly obscured by the brick molecular cloud. In such a case, the true size and the energy estimate are likely larger. Overall, we remark that, if these SNR candidates are real, their dynamical timescales are extremely short, which would imply an extremely high supernova rate. The SN rate in the CMZ has been estimated to he as high as 0.4 SN per millenium (Crocker et al. 2011). However, we caution that our dynamical time-scales could be off either because of a higher ambient density than we have assumed, or because absorption or confusion effects do not allow us to distinguish the proper border of the SNRs, or to detect any colder and more extended portions that might be present. Similar parameters characterise G0.52-0.046 (kT = 0.77 ± −3 0.3 keV, Apsho = 5+7 , NH = 7.9 ± 1.1 × 1022 cm−2 ). −3 × 10 Therefore, this feature also appears to be consistent with an SNR origin. However, although we derive a dynamical age of the same order (tdy ∼ 1700 yr), the energy inferred for the SN explosion is substantially lower, Eth ∼ 5 × 1049 erg. 8.3.2 The spectrum of G0.57-0.001 is characterised by significantly lower statistics and a higher column density of absorbing material. The best fit prefers a low temperature plasma, at the limit of detection. We fix its temperature to a relatively low value of −3 kT = 0.6 keV and find Apsho = 4.3+11 and NH = −4 × 10 22 −2 9.5 ± 2 × 10 cm . The derived thermal energy and dynamical times are Eth ∼ 2.6 × 1049 erg and tdy ∼ 1600 yr, respectively. The soft X-ray excess of this feature lies spatially very close to a region of X-ray excess that has been traced by Fe XXV line emission (Nobukawa et al. 2008). We also note that the region defining G0.57-0.001 almost completely contains a diffuse X-ray source detected both by the Chandra and ASCA satellites (Senda et al. 2002). The Chandra image shows a very compact (∼ 10” radius) and hot shell, G0.57-0.018, possibly the youngest SNR in the Galaxy, less than about 100 yr old. However, Renaud et al. (2006), because they c 0000 RAS, MNRAS 000, 1–?? Expanding molecular shells The observed temperatures, ages and sizes of these SNR candidates are consistent with a Sedov-Taylor framework expanding into an average ambient density between 1 − 10 cm−3 (higher density environments would result in older and cooler SNRs; Ostriker & McKee 1988). If the X-ray enhancements described above truly arise from SNRs interacting with and carving bubbles inside or near the surfaces of molecular clouds (preceded, perhaps, by the wind of the massive progenitor), we should observe clear traces of such events also in the kinematics of the surrounding molecular matter. Such a complex and delicate investigation is beyond the scope of the present paper. Nevertheless, we note that Tanaka et al. (2009) discovered an expanding SiO shell (SiO0.56-0.01) centered at l ∼ 0.56◦ , b ∼ −0.01◦ and having a size of ∼ 3.0 × 3.4 pc2 . The center and size of the expanding SiO shell closely match the peak and size of the X-ray emission of G0.57-0.001 and suggest an association between the two. In particular, high-velocity clumps have been found consistent with the idea that the SiO shell consists of swept-up material. Tanaka et al. (2009) calculated a kinetic energy of Ekin ∼ 1050.4 erg for SiO0.56-0.01. This strongly suggests that 26 G. Ponti et al. parameter kT Apsho τ †‡ Size NH χ2 /dof ne tdy Eth unit keV s cm−3 pc 1022 cm−2 cm−3 yr erg G0.40-0.02 G0.52-0.046 G0.570-0.001 G0.224-0.032 Back 0.55 ± 0.1 −2 2.4+5 −1 × 10 > 1.7 × 1011 8.6 × 5.5 7.7 ± 0.8 807/777 0.77+0.7 −0.2 +9 5−3 × 10−3 > 7.5 × 1010 5.9 × 2.7 7.9 ± 1.1 342/364 0.6‡ −3 4.3+11 −4 × 10 > 1.4 × 1010 4.2 × 2.1 9.5 ± 2 141/134 0.54 ± 0.1 −2 3+6 −2 × 10 > 3.5 × 1011 5.4 × 2.6 7.4 ± 1 369/338 0.55† < 6 × 10−3 7.5 × 1011 † 8.6 × 5.5 8.8 ± 1 834/741 1.4 3.7 × 103 1.9 × 1050 1.4 1.7 × 103 5.0 × 1049 2.1 1.6 × 103 2.6 × 1049 11 1.8 × 103 2.6 × 1050 Table 7. Best-fit and derived parameters of the SNR candidates described in § 8.3. †Parameter unconstrained. ‡Value weakly constrained by the high column density of neutral material, therefore fixed for the corresponding fit. †‡ Ionisation time-scale of the shock plasma model. G0.57-0.001 is indeed a SNR caught in the process of carving its bubble. Further studies of the gas kinematics around the other Xray enhancements are required to establish their real nature. We note that the thermal energy estimated for these SNRs is observed to be systematically lower than the theoretical value for the remnant of a standard type II SN expanding into the interstellar medium. This might result from a relatively higher ambient density in the GC, leading to greater energy dissipation, or from a significant fraction of the energy budget going into the inflation of the bubbles and the production of cosmic rays. We also note that G0.570-0.001, G0.52-0.046 and G0.40-0.02 are located within the trail of massive stars that have been hypothesised to have tidally escaped from the Quintuplet cluster (see Fig. 17 and Habibi et al. 2013; 2014). This raises the possibility that some of these SNRs might be associated with SN explosions from stars originating in this massive, young stellar cluster. 8.4 Origin of the Sgr A X-ray lobes All the soft X-ray maps (see Figs. 3, 10, 11, 7, 12 and 14) show the presence of two extended features, with a size of roughly 5 − 10 pc, located to the Galactic north and south of Sgr A? , the socalled ”bipolar Sgr A lobes” (Morris et al. 2003; 2004; Markoff et al. 2010; Heard & Warwick 2013). 8.4.1 Lobe morphology The lobes appear to have roughly oval shapes with co-aligned major axes oriented perpendicular to the Galactic plane. They appear joined at the position of Sgr A? suggesting the latter is their point of origin (see Fig. 8). The top panel of Fig. 11 and Fig. 7 show that the lobes’ emission is characterised by a smaller ratio of soft X-ray lines to continuum (therefore characterised by a greener colour) compared to the surrounding regions (appearing with a redder colour) such as the superbubble, G0.1-0.1 and the Radio Arc (Fig. 11, 7 and 8). This suggests that the lobes, although they show thermal emission lines (see §4.2), have either a stronger nonthermal component or significantly hotter thermal emission than the surrounding regions10 . We also note that the eastern portion of what appears to be part of the southern lobe has a colour as red as 10 A hot plasma, with temperatures of ∼ 2 − 4 keV, produces intense Xray emission but weaker soft X-ray lines, compared to a plasma having a temperature around 1 keV. G0.1-0.1 and the superbubble regions (see Fig. 10). Therefore, this emission might not be associated with the lobes, but rather with G0.1-0.1 or the edge of the superbubble (however, a gap such as might be produced by a foreground dust lane, appears to separate the lobes’ emission from G0.1-0.1). Figures 7 and 8 also show that the surface brightness of the northern lobe decreases with distance from Sgr A? (Heard & Warwick 2013). The bottom panel of Fig. 11 shows that the lobes have an orange colour, indicating harder soft X-ray emission compared to the surrounding regions. In particular, a brighter and harder (yellowgreen) linear structure outlines the northern lobe, shaping it to have a well-defined and symmetric spade structure, with a sharp transition at the border. The sharpness of the transition suggests the presence of a limb-brightened shock, indicating that the lobe is a bubble enclosed by a thin shell of hot, compressed material. This claim is strengthened by our analysis of the Chandra data (see Fig. 8 and 9, Baganoff et al. 2003; Lu et al. 2008; Muno et al. 2008). The superior Chandra spatial resolution, in fact, allows us to note that these projected linear features are running right along the lobes’ edges, indeed confirming the presence of a shock (Fig. 8 and 9). We also note that the emission from the northern half of the lobe seems to be mainly due to three harder filaments converging at the top in a cusp, having radio continuum counterparts (Zhao et al. 2015) and associated Paschen-α emission, indicating that these are thermal features (see Fig. 18 and 19). In the southern lobe, two bright knots are observed in the center and at the tip. Interestingly these appear to be located approximately at the same distance and in the opposite direction, compared to Sgr A? , as two enhancements present in the northern lobe. Moreover, the two bright knots have a green-yellow colour (upper panel of Fig. 11) similar to their apparent counterparts in the northern lobe. This suggests both: i) a similar physical origin for these features in both the north and south lobes and; ii) that the process that created the lobes is symmetric about the Galactic plane and its engine is (or is located close to) Sgr A? . The obvious interpretation of this morphology is that energetic events simultaneously ejected diametrically opposed blobs of hot gas. However, upon closer inspection, the northern and southern lobes do not appear completely symmetric. For example, compared to the northern lobe, the western side of the southern lobe and the region close to Sgr A? appear suppressed (see Fig. 7, 8 and discussion following). On the other hand, the eastern side of the southern lobe appears to extend further (e.g. further east compared to G359.977-0.076) than the corresponding boundary of the northern lobe (located close to e.g. G359.974-0.000). As described before, c 0000 RAS, MNRAS 000, 1–?? 27 the emission around and east of G359.977-0.076 has a different colour and might therefore be associated with either the superbubble, with G0.1-0.1, or it could be a feature that is independent of either of these and of the lobes. The bottom panel of Fig. 15 shows that the region with depressed soft X-ray emission south of Sgr A? and on the western side of the southern lobe spatially coincides with the presence of the 20 km s−1 molecular cloud, which is thought to be located in front of Sgr A? (Coil et al. 2000; Ferrière 2009) and to have a large column density. Soft X-ray emission could be produced there but be completely obscured to us by this intervening cloud (see Fig. 2). To reinforce this idea, we note that, in fact, at this position, hard X-ray radiation (between 4.5 and 12 keV) is observed by Chandra to have a non-thermal spectrum and to be extended (Morris et al. 2003). Therefore, this hard radiation, which is able to penetrate the cloud, could be produced by strong shocks at the bubble’s border. Similar hard non-thermal filaments are observed in several places at the border of the northern lobe (Morris et al. 2003). This strongly suggests that the actual border of the southern lobe is located further west than the images reveal, and that the lobe’s soft X-ray emission is obscured there. Therefore, once the effect of absorption by molecular clouds (e.g. the 20 km s−1 cloud) is considered, the symmetry between the northern and southern lobes appears more clearly. The linear or filamentary structures (such as G359.974-0.000, G359.970-0.009, G359.959-0.027, G359.9450.044, G359.942-0.045, 359.933-0.039, see Fig. 8) observed in the northern lobe might be present in the southern one as well, but be suppressed by the intervening absorption. Figure 18 shows the comparison between the Chandra and Paα emission. The Paα map clearly show the presence of tendrils of foreground absorbing material, running north-south along the extension of the lobe. Interestingly, the same tendrils are also evident as absorption lanes in the X-ray image. We also note that two Paα emitting luminous stars are contributing to, if not dominating, the ionisation of the gas surrounding them. However, there is a close correspondence between the Paα emission and the soft Xray emission at the northern tip of the lobe, so strong shocks might be contributing both to the ionization and to the production of very hot post-shock gas that can emit X-rays. Indeed, we hypothesize that some portion of the hot wind that we see in the softest X-ray bands, presumably emanating from near Sgr A? , is undergoing a shock where it encounters ambient interstellar material at b ∼ 0.02 degrees, and that it is thereby blocked from continuing to higher latitudes. That shock is manifested both as a horizontal feature in the Paα image, and as a diminution in the brightness of the X-ray emission proceeding north from that latitude. There is still some weaker, extended X-ray emission at higher latitudes, indicating that not all of the outflowing wind is completely blocked. In addition, the portion of the X-ray emitting plasma lying behind the shock front appears to be absorbed along the shock front, creating a horizontal shadow in the extended X-ray emission, and suggesting that the shock at b ∼ 0.02 degrees has created a thick, compressed layer that absorbs the X-rays coming from behind it. 8.4.2 Lobes collimated by the Circumnuclear Disk (CND) Morris et al. (2003) observed that the CND has a size and orientation that are consistent with it being the agent that collimates an isotropic outflow from the Sgr A? region, thereby creating the bipolar lobes of hot plasma. Those authors further suggested that the sequence of enhancements along the axis of the lobes might have resulted from a series of energetic mass ejections from the immec 0000 RAS, MNRAS 000, 1–?? diate environment of Sgr A? . A similar scenario has recently been discussed by Heard & Warwick (2013b). At an outflow velocity of 103 km s−1 , it would have taken 2 − 6 × 103 yr to inflate the lobes. Assuming a thermal emission model, an electron density ne in the range 1 − 10 cm−3 can be inferred (Morris et al. 2003; Heard & Warwick 2013b) and, assuming that the line-of-sight depth is equal to the projected width, the total hot plasma mass involved in the X-ray lobes is only about 1 − 3 Solar masses. 8.4.3 Energetics of the Lobes Except for the broad intensity enhancements along the axis of the lobes, the surface brightness is relatively smoothly distributed. Assuming a continuous and constant outflow, we examine the energy budget of the lobes. In particular, integrating the measured energy density over a cylinder of 5 pc radius and 12 pc height (the approximate sizes of the lobes), a thermal energy of Eth ∼ 9 × 1049 erg is estimated. It has been estimated that the massive stars in the central parsec collectively lose ∼ 5 × 10−3 M yr−1 in stellar winds, with velocities ranging from ∼ 300 to 1000 km s−1 (Geballe et al. 1987; Najarro et al. 1997; Paumard et al. 2001). The total kinetic energy thermalised by shocks is then E ∼ 5×1038 erg s−1 for such a mass outflow rate and an outflow velocity of 1000 km s−1 (Quataert & Loeb 2005). Therefore the energy released within the time needed to inflate the lobes (∼ 4×103 yr) is equivalent to E ∼ 5×1049 erg, therefore giving an important contribution to the generation of the lobes. The lobes might also be traceable to the accretion flow onto Sgr A? . We note that, as estimated by Wang et al. (2013), only ∼ 1 % of the matter initially captured at the Bondi radius presently reaches the innermost regions around Sgr A? . The rest of the accretion power, estimated to be ∼ 1039 erg s−1 (Wang et al. 2013), is probably converted to kinetic energy and used to drive an outflow that carries away the bulk of the inflow, sculpting the environment with its ram pressure. If so, within the lobe inflation time, a total energy of ∼ 1050 erg would have been deposited. Within the CMZ, X-ray reflection nebulae indicate that a few hundred years ago Sgr A? was more active, being ∼ 106 times brighter than at present for approximately 5 − 10 % of the time in the past millennium (experiencing LX ∼ 1039 erg s−1 ; see Ponti et al. 2013 for a review). Could the lobes have been created by similar events that occurred over the past 104 yr? The light crossing time of the CMZ limits our capability to trace Sgr A? ’s past activity beyond about 103 yr ago, so it is difficult to directly trace the echoes of possible energetic events on such time scales. However, if the process has been active over the past (5 − 10) × 103 yr at roughly the same rate, (therefore active at L ∼ 2 × 1039 erg s−1 for ∼ 103 yr in the past 104 yr), a total integrated energy equal to ∼ 5×1049 erg should have been generated. If Sgr A? ’s past activity was characterised by outbursts with associated outflows having kinetic luminosity comparable to the radiated power (therefore much higher than in soft state stellar mass black holes; Ponti et al. 2012), then these events could be the primary source (or at least contribute) to form the lobes. All these processes appear similarly likely to have an impact on the formation of the lobes, from an energetic point of view. However, we note that the first two mechanisms are powered by a quasi-continuous outflow from Sgr A? or from the central stellar cluster. In such scenarios, therefore, the sharpness of the edges at the extremities of the lobes remains rather puzzling, favouring explosive-outbursting scenarios. 28 G. Ponti et al. Figure 18. (Left panel) Chandra RGB image of the northern lobe. See Figure 8 for more details. (Right panel) Pa-α image of the northern lobe, tracing thermal, ionized gas (from Wang et al. 2010). Continuum emission from stars has been removed (Dong et al. 2011), so the only stars that appear are those that have strong Pa-α emission lines. Two Paα-emitting luminous stars located at l ∼ 359.935◦ , b ∼ 0.21◦ and l ∼ 359.925◦ , b ∼ 0.45◦ (Mauerhan et al. 2010; Dong et al. 2012) are probably responsible for at least part of the ionisation indicated by the Pa-α. See WWW. MPE . MPG . DE / HEG / GC / for a higher resolution version of these figures. 8.4.4 Are Sgr A’s lobes the SNR of SGR J1745-2900? The recent discovery of a young magnetar, SGR J1745-2900 (Degenaar et al. 2013; Dwelly & Ponti 2013; Mori et al. 2013), most probably in orbit around Sgr A? (Rea et al. 2013), raises the quest for finding its young SNR. SGR J1745-2900 is estimated to be only about 9 × 103 yr old and to be located at < ∼0.07 − 2 pc from Sgr A? (Rea et al. 2013). Therefore, the supernova that generated SGR J1745-2900 should have exploded near the centroid of the lobes and likely inside the inner radius of the CND. If the shock propagated at the sound speed (vs ∼ 750 km s−1 ), then a present size of ∼ 6.5 pc would be expected. This is slightly smaller than the observed size of the lobes (∼ 12 pc), but is consistent with the observed size within the uncertainties in the age and the sound speed. We also note that the thermal energy content in the lobes (Eth ∼ 9 × 1049 erg) is lower, but of comparable order of magnitude, to the energy released by typical supernova explosions. It is therefore plausible that the lobes are indeed the SNR associated with SGR J1745-2900. Another viable possibility is that the lobes have been generated by the supernova that created the PWN candidate G359.9450.044, located only ∼ 8” from Sgr A? , with an estimated age of few thousand years (Wang et al. 2006). It is therefore possible that at least one SN exploded close to Sgr A? within the last ∼ 10 kyr. If so, its blast wave likely propagated into a pre-existing hot, low density cavity created by Sgr A? ’s outflows and the collective winds of the central stellar cluster. Given the density and the temperature in the lobes (kT ∼ 2 keV; Morris et al. 2003), the shock is estimated to reach 15 pc in 9 × 103 yr (Wang et al. 2005), a value consistent with the observed size of the lobes. As mentioned above, the presence of the sharp edges to the lobes seems to favour an explosive mechanism for their creation. A supernova exploding inside the CND would expand into the preexisting, stationary outflow from the center and be collimated in the same way. Hydrodynamical simulations typically show that SN shock fronts are reflected away when encountering the walls of a dense molecular cloud, such as the CND (Ferreira & de Jager 2008). In this scenario the sharp edge of the lobes would be due to the SN shock front. If the supernova recurrence time is longer than or comparable to the lobe expansion time (a few thousand years) then this would not appear as a stationary process. Assuming a recurrence time between 1−10×104 yr (similar to the SN recurrence time of the central young cluster, in which ∼ 100 massive stars presumably becoming SN over a ∼ 107 yr time interval), we estimate a time-averaged kinetic power release of 3−30×1038 erg s−1 . This indeed suggests that: i) SN explosions of the central star cluster can c 0000 RAS, MNRAS 000, 1–?? 29 contribute to powering the lobes; ii) the lobes are quasi-stationary features; and iii) it is not unlikely that we observe such features created by a rare event such as a SN explosion. Finally, we note that, although the characteristics of the Xray emission from the lobes appear consistent with being the X-ray remnant of a recent supernova that exploded within a few parsecs of Sgr A? , the lack of associated nonthermal radio emission from such a young SNR is problematical for this hypothesis. Mori et al. (2009) and Heard & Warwick (2013) estimate a total thermal energy contained within the superbubble of Eth ∼ 1051 f 1/2 erg (where f is the volume filling factor of the emitting plasma). Such a large energy content does, indeed, require multiple supernova events. Those authors also estimate for G359.77-0.09 and G359.79-0.26 an ionisation time-scale of tion ∼ 3 × 104 yr (assuming f ∼ 1). 8.5.2 8.5 G359.77-0.09 and G359.79-026: a ring from a hot superbubble southwest of Sgr A? A series of diffuse, soft features appear to the southwest of Sgr A? (see Fig. 3), namely G359.77-0.09, G359.79-0.26 and a newly recognised extended feature, G0.0-0.16. If these are physically connected, they form, in projection, a roughly elliptical shape whose major axis has an inclination of about 60◦ with respect to the Galactic plane (see Fig. 3, 6, 7, 10, 11, 12 and 14). These features show similar colours and strong, soft line emissions, indicating a similar thermal origin (see Fig. 3, 10 and 11). This elliptical structure appears brightest at the softest energies, however it is not observed below 1.5 keV, suggesting a location near the Galactic centre and a low temperature for its plasma, compared to the surrounding emission. We note that this feature is characterised by a very bright edge with strong Si XIII emission on the outside of the ellipse, suggesting a lower temperature of the edge compared to the interior. The ellipse center is located at l ∼ 359.9◦ , b ∼ −0.125◦ and it has minor and major axes of about 7.8 and ∼ 12 arcmin, respectively (corresponding to 18 and 28 pc). Both Mori et al. (2009) and Heard & Warwick (2013) considered that these structures/group of structures were physically connected and form a superbubble candidate. We note that the recognition of such an elliptical ring critically depends on the presence of the dark lane running from l ∼ 0.02◦ , b ∼ −0.22◦ to l ∼ 0.05◦ , b ∼ −0.07◦ (see Fig. 3, 10, 11, 7 and 14), which separates G0.0-0.16 from the emission of G0.1-0.1; this dark lane helps define the quasi-continuous elliptical morphology of the ring. However, the lane might simply be due to absorption by foreground material, in which case G0.0-0.16, forming the eastern part of the ring, could simply be connected to G0.1-0.1. 8.5.1 S XV emission filling the superbubble The S XV emission provides a key piece of information to better understand the superbubble. We note, in fact, that the S XV emission completely fills the region inside the ring with a roughly uniform brightness (several times brighter than in the surrounding region) and sharply drops at the ring’s edge. This indicates that the superbubble is a shell of hot gas that we see projected onto the plane of the sky (see middle top panel of Fig. 12). To further corroborate this, we note that a sharp emissivity drop appears to be located just outside of the ring, running all around the ring’s external edge. Such X-ray depression might be produced by a high column density of cold gas pushed away by the superbubble’s shock front and accumulated in large quantities just outside the shock. If so, the observed depression could be indicating that the superbubble is located in front of G0.1-0.1. Fig. 7 also shows a small depression in the top part of the northern lobe that could easily be explained by absorption associated with the superbubble if it is located in front of the lobes. This situation would then be somewhat analogous to the colour variation in the south lobe. c 0000 RAS, MNRAS 000, 1–?? Origin The origin of such a superbubble is not clear. We note that many of the massive stars that are suggested to have escaped from the Quintuplet cluster (Habibi et al. 2013; 2014) are projected inside the superbubble. It is therefore possible that explosions of stars lost by the Quintuplet cluster have contributed to energising the superbubble. A more speculative point is that the estimated age of the superbubble is of the same order of magnitude as the recurrence time of tidal disruption events by Sgr A? : tT DE ∼ 1 − 3 × 104 yr (Alexander & Hopman 2003). While Sgr A? appears, in projection, to be located inside the superbubble, it is ∼ 6.7 arcmin off from the superbubble’s center. Khokhlov & Melia (1996) suggested that an explosion associated with a tidal disruption event would liberate a large amount of energy on the order of E ∼ 1052 erg that would propagate as a powerful shock wave into the local interstellar medium. As with the remnant of the SGR J1745-2900, we expect that the shockwaves of a tidal disruption event would interact with the CND. However, the unbound part of the tidally disrupted star would be ejected into a limited solid angle, producing a strongly elongated and asymmetric remnant (Khokhlov et al. 1996; Ayal et al. 2000). Such a remnant would then appear as a very energetic shell of hot gas and remain visible for a time comparable to the age of a typical SNR. Assuming a shock survival time of ∼ 1 − 10 × 104 yr, we could potentially observe a few remnants resulting from tidal disruptions. We suggest that the superbubble G359.77-0.09 has properties that make it a possible candidate. No other feature with properties obviously related to a tidal disruption event appear to be observed close to Sgr A? . 8.6 The arc bubble: a second superbubble in the GC Highly enhanced soft X-ray emission is observed east of Sgr A? from the region called G0.1-0.1. This feature appears in Figures 10 and 11 as a slightly elliptical feature of enhanced emission with center at l ∼ 0.09◦ , b ∼ −0.09◦ and with radius of ∼ 5 arcmin (corresponding to ∼ 10 pc). The top panel of Fig. 11 shows that G0.1-0.1 and the Radio Arc regions both show distinct red emission. This indicates large equivalent widths of the emission lines from this plasma and therefore a thermal origin. However, the top panel of Fig. 10 and the bottom of Fig. 11 show strong colour gradients within these regions, indicating that they might have different contributions from distinct components. The PWN candidate G0.13-0.11 (l = 0.131◦ , b = −0.111◦ ; Wang et al. 2002; Heard & Warwick 2013) stands out from the general thermal emission in G0.1-0.1, appearing as a distinct point-like source characterised by a light blue/white colour11 , indicating its non-thermal origin (dominated by intense soft and hard continuum 11 Note that in the top panel of Fig. 10 despite an enhancement of diffuse emission from the region surrounding the core of G0.13-0.11, no point like source is detected in the soft X-ray line image. 30 G. Ponti et al. emission with no soft line emission, see the top panel of Fig. 11)12 . The point-like head of G0.13-0.11 appears to be accompanied by a tail extending to the south for 4.5 − 5 pc; in Fig. 19, it appears with a white-violet colour. Heard & Warwick (2013) suggest that the SN that generated G0.13-0.11 might be the source of the soft X-ray emission from this region. Those authors present a spectral study of the X-ray emission from G0.1-0.1 and find a gas temperature of kT = 1.1 ± 0.1 keV, and a column density of NH = 5.6 ± 0.5 × 1022 cm−2 , indicating a GC location of this emission, and abundances that are about 1.8 times Solar. Assuming that the plasma volume is only 3.5 pc3 , corresponding to only 1.5 arcmin radius around the PWN (see the red circle in Fig. 20), the authors estimated a thermal energy of Eth = 3.1 × 1049 erg (and a plasma ionisation time-scale of at least t = 1.8 × 104 yr), thus consistent with being produced by a single supernova explosion. 8.6.1 Energetics of the arc bubble We note from Figs. 19, 3, 10, 11, 7, 12 and 14 that G0.1-0.1 extends further from G0.13-0.11 than the 1.5 arcmin region size considered by Heard & Warwick (2013), with no clear boundary at 1.5 arcmin (see Fig. 20). To illustrate this, Fig. 19 shows the soft emission lines RGB image with colour scales chosen to highlight intensity variations present within this region. The left panel of Fig. 20 shows the contours of the S XV emission overlaid on the 20 µm MSX image (Price et al. 2001). Figure 20 clearly shows that the empty mid-IR bubble (the so called arc bubble; Levine et al. 1999; RodriguezFernandez et al. 2001; Simpson et al. 2007) is completely filled with warm X-ray emitting plasma and that the soft X-ray emission is not confined to within ∼ 1.5 arcmin of G0.13-0.11, but it extends much further, for about 7 arcmin. Assuming a uniform surface brightness, if the bubble is 4.5 times larger, we would expect a thermal energy of Eth ∼ 1.5 × 1051 erg, thus most probably requiring multiple supernova events, and supporting the notion that G0.1-0.1 is a second superbubble candidate in the GC. Rodriguez Fernandez et al. (2001) have noticed that the radio Arc bubble is filled with continuum X-ray emission seen by ASCA which they ascribed to X-ray sources inside the bubble. Here we find that the bubble is in fact filled with diffuse thermal X-rays, most likely originating from SN explosions of massive stars associated with the Quintuplet cluster. We also note that the soft line emission is at least as extended as the arc bubble and is highly inhomogeneous (see Figs. 19 and 20). Three depressions having roughly circular shapes can be discerned in G0.1-0.1 (Fig. 20). Two cavities are located at about the same latitude, with centers close to l = 0.057◦ , b = −0.067◦ and to l = 0.116◦ , b = −0.071◦ and with radii of ∼ 1.6 arcmin (corresponding to ∼ 3.7 pc) and ∼ 1 arcmin, respectively. These cavities appear to be surrounded by a thin rim of brighter emission. A third depression is centered at l = 0.083◦ , b = −0.123◦ , with ∼ 1.8 arcmin radius. This cavity also seems to be confined by a thin shell of brighter material, except for its southern edge, where it appears open (see Fig. 19), possibly because of the presence of a dark absorbing lane. 12 Please note that an unrelated point source at l = 0.142◦ , b = −0.109◦ is located at the same latitude, but ∼ 1.5 pc to the Galactic east of the PWN candidate G0.13-0.11. 8.6.2 Association with the Quintuplet cluster? Despite its offset position from the centre of the mid-IR arc bubble, the Quintuplet cluster is often considered responsible for creating and maintaining the bubble with a combination of supernovae and strong stellar winds (e.g., Simpson et al. 2007). Johnson et al. (2007) invoke a possible non-uniformity of the ambient medium as a possible origin of this asymmetry. This might also apply to the Xray emission. However, we note that the Quintuplet cluster is moving supersonically within the CMZ (Stolte et al. 2014). Given its projected velocity, the cluster would have taken ∼ 100 kyr to cross the width of the IR arc bubble, in which case the bubble would have been inflated on a time scale much smaller than typical superbubble formation times (Castor et al. 1975; Weaver et al. 1977; Mc Low & McCray 1988). Several SN explosions in that amount of time would therefore have been required. Furthermore, the Quintuplet cluster would have been located in the middle of the two northern cavities about 4 × 104 and 9 × 104 yr ago, respectively (the right panel of Fig. 20 illustrates the direction of the cluster’s motion and the position of the cavities). The relatively small cavities observed in the northern part of G0.1-0.1 are unlikely to have been generated by multiple cluster stars. In fact, a supernova exploding in the hot plasma of G0.1-0.1 is expected to undergo a significantly different evolution than a typical SNR. In particular, the sound velocity is significantly larger than in a typical low pressure medium. Tang & Wang (2005) have shown that the shock velocity follows a Sedov solution but quickly deviates from it when it becomes mildly supersonic. This translates into a much faster evolution and much larger cavities would be expected if the SN exploded 4 × 104 and 9×104 yr ago, when the quintuplet cluster was at that location. It is more likely, therefore, that the two cavities to the Galactic west of the Quintuplet might have been generated by supernova explosions of massive stars either stripped from the Quintuplet cluster (Habibi et al. 2013) or having no association with it. The large thermal energy filling the arc superbubble could have been produced by some combination of winds from the young stars and by multiple supernova explosions, including the supernova explosion associated with the PWN G0.13-0.1113 . 8.7 8.7.1 A hot atmosphere around the GC - A link to the GC lobe? And to the Fermi bubbles? General morphology As observed in all soft X-ray maps (Figs. 3, 10 and 11) and confirmed by the soft plasma intensity map (obtained through spectralimages decomposition, Fig. 14), the regions at higher Galactic latitudes are significantly brighter in soft X-rays than the regions closer to the disc, presumably in part because of the smaller extinction at the higher latitudes. The western border of this enhanced emission is defined by a relatively sharp edge between l = 359.63◦ , b = 0.06◦ and l = 359.55◦ , b = 0.46◦ (Fig. 14, see also Figs. 10 and 11). The soft X-ray emission peaks above the location of the GC Radio Arc, appearing as a continuation of the Radio Arc itself. Further west the soft X-ray emission appears to fade with increasing Galactic longitude. In particular, the spectral decomposition provides hints for the presence of an eastern edge, fainter but similar to the western edge, of this hot GC atmosphere. However the presence of an edge to 13 We note that the PWN G0.13-0.11 is located right in the middle of the mid-IR arc bubble. c 0000 RAS, MNRAS 000, 1–?? 31 0.150 0.100 0.050 Galactic latitude 0.000 -0.050 -0.100 -0.150 -0.200 -0.250 0.300 0.200 0.100 0.000 359.900 359.800 359.700 Galactic longitude Figure 19. RGB image with colour scale chosen to highlight enhancements and depressions in the diffuse emission east of Sgr A? . In red the sum of the Si-S and S XV bands is shown. In green the sum of S-Ar plus the S XV and Ar XVII bands is shown. The blue image shows the sum of the Ar-Ca plus Blue-Ca and Ca XIX bands (see Tab. 1 for the definition of the energy bands). the west is less obvious because of the soft X-ray extinction likely caused by a series of molecular clouds (e.g., the Brick) present at that location and because of the partial high-latitude coverage of this region. 8.7.2 Radiative process The presence of intense, soft X-ray emission lines (see Figs. 10 and 11) and, in particular, the good fit of a spectral decomposition based on a thermally emitting gas (see Fig. 14) indicate that most of the high-latitude emission is generated by a thermal radiative process in a warm plasma. To demonstrate this, we accumulate the EPICpn spectrum from a circular region of 8.28 arcmin radius centered at l = 0.181◦ , b = 0.359◦ . The resulting spectrum is well fitted with an absorbed thermal emission component (APEC) with kT = 0.96 ± 0.1 keV, NH = (2.3 ± 0.2) × 1022 cm−2 and Aapec = (1.5 ± 0.4) × 10−2 cm−5 . might have two contributions, one associated with the G0.1-0.1 superbubble (filling the mid-IR arc bubble; see § 8.6), while the second is associated with enhanced soft X-ray emission due to the presence of the Radio Arc and its polarized radio plumes at higher latitudes (Seiradakis et al. 1985; Tsuboi et al. 1986; Yusef-Zadeh & Morris 1988). If that is indeed the case and if the two structures have different X-ray colours (e.g. the superbubble produces lower temperature thermal X-ray lines, while the Radio Arc has a larger continuum to lines ratio), then we should observe variations in the X-ray colour distribution. In particular, we would expect a whiter colour and a green-yellow colour (similar to the one characterising the lobes of Sgr A) at the location of the Radio Arc region compared to G0.1-0.1, in the top panel of Fig. 10 and in Fig. 11, respectively. This idea is, indeed, in agreement with the colour variations and the evolution of the line intensities observed between the G0.1-0.1 and Radio Arc complexes (Fig. 10 and 11). 8.7.4 8.7.3 Eastern edge The eastern edge of the high-latitude emission rises from the position of the Radio Arc. This raises the interesting question of whether the soft X-ray emission at the location of the Radio Arc c 0000 RAS, MNRAS 000, 1–?? Western edge, the Chimney and AFGL5376 Running almost parallel to the western edge of the high-latitude plasma is another region of enhanced soft X-ray emission, located near the Galactic plane, the so called Chimney (l = 359.45◦ ; Tsuru et al. 2009). The Chimney appears as a column of soft X-ray emitting plasma extending all the way between the core of Sgr C and 32 G. Ponti et al. 0.100 0.100 North Lobe 0.050 0.050 Arches cluster 0.000 Galactic latitude Galactic latitude 0.000 -0.050 -0.100 PWN G0.13-0.11 -0.150 Quintuplet cluster -0.050 167 km/s Tail of massive stars -0.100 PWN G0.13-0.11 -0.150 South Lobe -0.200 -0.200 0.250 0.200 0.150 0.100 0.050 0.000 359.950 359.900 -0.250 0.250 0.200 0.150 0.100 0.050 0.000 359.950 359.900 -0.250 Galactic longitude Galactic longitude Figure 20. (Left panel) 20 µm MSX map of the GC. The contours indicate the intensity of S XV emission. Soft X-ray emission fills the arc bubble observed in the mid-IR. The green solid circle and the white dashed ellipses indicate the position of the PWN G0.13-0.11 and three structures in the soft X-ray emission map (see right panel). (Right panel) Soft X-ray map of the GC (the same energy bands used in Fig. 19 are displayed). The position of the PWN G0.13-0.11 is indicated by a red circle with 1.5 arcmin radius. At least three sub-structures, appearing like holes, are observed within G0.1-0.1 (here indicated with white dashed ellipses). The positions of the Arches and Quintuplet star clusters are indicated by yellow dashed circles. The direction of the supersonic motion of the Quintuplet cluster is indicated and its past location is indicated by the yellow dashed line. The inferred positions of the Quintuplet 4 × 104 and 9 × 104 years ago are indicated with yellow crosses. The cyan dashed ellipse indicates the region in which many massive stars that might have been expelled by the Quintuplet cluster are located. the northern limit of the XMM-Newton scan (b ∼ 0.15◦ ; see Fig. 10, 11 and 15). Tsuru et al. (2009) suggested that the Chimney is an outflow emanating from the supernova remnant candidate G359.41-0.12. They estimated a thermal energy and dynamical time for G359.41-0.12 and the Chimney of Eth = 5.9 × 1049 erg, tdy = 2.4 × 104 yr and Eth = 7.6 × 1049 erg, tdy = 4 × 104 yr, respectively. The energetics and time-scales are consistent with typical GC supernova remnants and Tsuru et al. suggested that the very peculiar morphology of the outflow producing the Chimney might be due to a peculiar distribution of molecular clouds that block the plasma expansion in the other directions (Tsuru et al. 2009). We note that the morphology of the Chimney resembles that of the Radio Arc. It originates near the Galactic plane (where dense and massive molecular clouds are located) and extends almost perpendicular to the Galactic plane. Within the gap between the Chimney and the western edge (see Fig. 6), a bright non-thermal radio filament with an X-ray counterpart is observed: G359.54+0.18 – the Ripple filament, with a radio length of 0.08◦ (Lu et al. 2003; YusefZadeh et al. 2005). It is oriented parallel to the edge of the soft Xray plasma distribution (Bally et al. 1989; Yusef-Zadeh et al. 1997; 2004; Staguhn et al. 1998; see also Yamauchi et al. 2014). Similar to the Radio Arc, other X-ray and non-thermal radio filaments are observed at the base of the Chimney. The high concentration of non-thermal filaments indicates the importance of magnetic structures in this region (e.g., Morris 2014). The bright IR source AFGL5376 is located further to the northwest, along the continuation of the sharp X-ray edge and the Chimney (unfortunately just off the XMM-Newton map; see Uchida et al. 1990; 1994). It is associated with high-velocity CO emission and defines the most prominent portion of a strong large scale (∼ 90 pc) shock front that extends all the way down to Sgr C (Uchida et al. 1994). Because the Chimney appears to be associated with a shock, and because it is spatially coincident with magnetic filaments along its length (Yusef-Zadeh et al. 2004), we suggest that it is not a simple supernova remnant, but is a phenomenon associated with a footpoint of a larger scale structure, the GCL. 8.7.5 Is the outflow confined inside the GCL? Inside the Fermi bubbles? The GC is considered a mini-starburst environment, producing intense outflows (Crocker 2012; Yoast-Hull et al. 2014). The warm plasma detected at high-latitudes is therefore, most probably, associated with intense star formation and it can be a pervasive atmosphere above the entire CMZ. Even in the absence of another confining force, the gravitational potential (e.g., that of Breitschwerdt et al. 1991; Launhardt et al. 2002) would bind the ∼ 1 keV plasma (having a sound speed of ∼ 500 km s−1 ; Muno et al. 2004) to the Galaxy, but in hydrostatic equilibrium, would allow it to extend to heights of several hundred parsecs. If so, it would require a large average star formation rate and concomitant energy input to generate and maintain it. However, the detection of edges in its distribution suggests that such plasma might be confined within known structures. As noted by Blanton (2008), the locations of the eastern and western footpoints of the GCL (Law 2011) correspond to the positions of the Radio Arc and the Chimney, respectively. Indeed, the GCL and its possible magnetic nature might confine the warm plasma observed in soft X-rays. This opens the exciting possibility that the observed high-latitude enhanced X-ray emission from the GC ”atmosphere” is indeed the warm plasma filling the GCL. Based on the spectral fit, we deduce a density of ne = 0.06 cm−3 inside the GCL. Assuming uniform physical conditions inside the GCL (a cylinder of 45 pc radius and 160 pc height) and extrapolating over the entire GCL, we estimate a mass of ∼ 4 × 103 M filling the GCL with a total thermal energy of ∼ 1052 erg. This value is of the c 0000 RAS, MNRAS 000, 1–?? 33 same order of magnitude as the energy required to inflate the GCL as estimated by Law (2011). Just after the first detection of the GCL, Uchida et al. (1985) noted its similarity with the lobes in nearby radio galaxies (although smaller in size and strength). The authors interpreted the lobes as created by a magneto-dynamic acceleration mechanism where the magnetic twist is produced by the rotation of a contracting disc of gas in the Galactic plane. Under such conditions, the plasma is accelerated into a conical cylinder with a helical velocity field (Uchida et al. 1985). Alternatively, Bland-Hawthorn & Cohen (2003) suggested that the GCL could be produced by a largescale bipolar Galactic wind, that would be the result of a powerful (E = 1054−55 erg) nuclear starburst that took place a few 106 yr ago. These authors show that dust is associated with the entire GCL structure and they suggest that the GC (and the centers of many Galaxies) would drive large-scale winds into the halo with a recurrence time of about 10 Myr (Bland-Hawthorn & Cohen 2003). Other, alternative, scenarios for the origin of the GCL involve outflows associated with enhanced activity of Sgr A? (Ponti et al. 2013) or intense star formation (Crocker et al. 2011; 2012). It is not excluded that the GCL could be simply one part of an even larger scale feature extending over a physical scale of several kiloparsecs above and below the Galaxy, the so called Fermi bubbles (Su et al. 2010). These gamma-ray bubbles, detected with Fermi, are interpreted as produced by highly relativistic particles emitting brightly at GeV energies and beyond and they appear to contain and confine soft X-ray emitting plasma traced by the ROSAT all sky survey, from very large scales down to the Milky Way’s center (Su et al. 2010). However, close to the Galactic plane, the bubbles’ edges start to become confused. Whatever their origin might be, the Fermi bubbles appear to originate (and be collimated) from the CMZ, within the region that XMM-Newton scanned here. Additional XMM-Newton observations at high Galactic latitudes, in particular, inside and at the border of the GCL, covering the AFGL 5376 source and the edges of the base of the Fermi bubbles will be needed to measure the extent of this high-latitude emission and to help disentangle the hypotheses for its origin. Furthermore, higher spatial resolution observations (such as provided by Chandra) at high latitudes would allow one to pin down what fraction of the extended, high-latitude X-ray emission is associated with faint point sources that are relatively less subject to extinction than sources near the plane. 8.8 Soft X-ray emission from the Sgr D and Sgr E regions Intense, soft, diffuse X-ray emission is observed from G359.120.05, the region around 1E1740.7-2942. A radio SNR (G359.070.02) is observed at about the same position (LaRosa et al. 2000). G359.12-0.05 has an emission spectrum typical of an SNR (Nakashima et al. 2010). In particular, the high extinction suggests it is located at the GC. Nakashima et al. (2010) suggest that G359.12-0.05 might be associated with the great annihilator and therefore be the second system (such as SS433 and the radio SNR W 50) where a BH is associated with its SNR. The core of the Sgr D complex is also observed to show enhanced medium energy emission (see Fig. 3). In radio, a SNR southwest of Sgr D’s core and H II regions are clearly observed (Fig. 6; LaRosa et al. 2000). Sawada et al. (2009) analysed X-ray data from the XMM-Newton and Suzaku satellites and observed soft X-ray emission from two diffuse X-ray sources, DS1 and one associated with the core of Sgr D. They suggest that DS1 is a new SNR at the GC. c 0000 RAS, MNRAS 000, 1–?? 8.9 Star formation estimate from counts of supernova remnants We observe a total of ∼ 10 − 12 supernova remnant candidates in the CMZ (plus ∼ 5 independent radio SNR; see Tab. 7) plus two superbubbles, each likely created by many (3-10) supernova events. These remnants have typical estimated ages of a few tens of thousands of years and temperatures of kT ∼ 0.4 − 1.5 keV. Due to the presence of the two superbubbles requiring multiple SNe and the high absorption towards the GC (such as in the star forming region Sgr B) that hampers us from observing a potentially larger population of remnants characterised by lower temperatures, the number of SNR observed in the GC is most probably under-estimated. However, assuming lifetimes of 10-40 kyr, the observed number of SNR yields a rate, averaged over the past several thousands of years, of rSN ∼ 3.5 − 15 × 10−4 yr−1 , consistent with other estimates (Crocker 2012). This implies a kinetic energy input higher than 1.1 × 1040 erg s−1 . To estimate the star formation rate, we assume that all stars with masses greater than 8 M produce supernovae and that all SNR are observable. Therefore, we multiply the supernova rate by the integral of the initial mass function (IMF) over all masses divided by the integral of the IMF above 8 M . To reflect the GC IMF, we assume the Kroupa (2002) formulation. The star formation rate then results to be: rSFR ∼ 0.035 − 0.15 M yr−1 . If the IMF in the CMZ is top-heavy, as some have argued, then a smaller star formation rate is implied. As noted also by Mori et al. (2008; 2009) and Heard & Warwick (2013), we observe that the two superbubbles have far hotter temperatures (higher density and smaller size) than all the ones observed in the Galactic plane or in the Large Magellanic Cloud (typically with temperatures kT ∼ 0.1 − 0.3 keV, densities ne ∼ 0.01 − 0.03 cm−2 , sizes l ∼ 140 − 450 pc; but see also Sasaki et al. 2011; Kavanagh et al. 2012). This could simply be the consequence of the high extinction towards the GC, hiding a population of normal and very soft superbubbles, or it could be a characteristic feature of GC superbubbles, inducing a different evolution because of the interaction with the peculiar GC environment. Further investigation is required to solve this problem. 9 CONCLUSIONS We have systematically analysed more than 100 XMM-Newton observations pointed within one degree of Sgr A? and have created the deepest, few arcsec resolution, X-ray images of the CMZ. This includes a total of about 1.5 Ms of EPIC-pn cleaned exposure in the central 15 arcmin and about 200 ks at all other points of the Central Molecular Zone (CMZ). We present here, for the first time, not only broad-band X-ray continuum maps, but also mosaicked maps of both soft line intensities and inter-line emission from the entire CMZ region. • The remarkably similar distributions of both the soft line emitting plasma (Si XIII, S XV, Ar XVII and Ca XIX) and the soft continuum (intra-line bands) indicate that most of the diffuse soft X-ray emission arises from a thermal process generating both continuum and lines. • Starting from the mosaic maps of the different narrow energy bands and assuming the GC emission is produced by three different components, we fit the maps at different energies and derive the integrated intensity map of the thermal plasma emission. Integrating over the entire CMZ, the total observed (un- 34 G. Ponti et al. absorbed) flux is: F2−4.5 keV = 4.2 × 10−11 erg cm−2 s−1 , F4.5−12 keV = 1.2 × 10−10 erg cm−2 s−1 , corresponding to a luminosity of L2−12 = 3.4 × 1036 erg s−1 at an assumed 8 kpc distance. • Counting the number of supernova remnants in the CMZ, we estimate a supernova rate between rSN ∼ 3.5 − 15 × 10−4 yr−1 , consistent with other estimates (Crocker 2012), that corresponds to a star formation rate of rSFR ∼ 0.035 − 0.15 M yr−1 over the past several thousand years. This implies a kinetic energy input greater than 1.5 × 1040 erg s−1 . • We report the discovery of a new X-ray filament XMM J0.173-0.413 perpendicular to the Galactic plane and south of the GC Radio Arc spatially corresponding to a nonthermal radio filament. XMM J0.173-0.413 is the first of the four cases known where the X-ray emission is not at or near a location where the radio filaments show unusually strong curvature. • The soft GC X-ray emission is absorbed not only by highcolumn-density foreground clouds located in the Galactic disk, but also by some clouds located on the near side of the CMZ, such as the core and envelope of Sgr B2, M0.25+0.01 (the ”Brick”), and even a few clouds at higher Galactic latitudes, M0.18+0.126 and M0.20-0.48. However, the majority of the observed variations in the soft X-ray emission are true emissivity modulations and not a product of absorption. • Several SNR candidates are identified by their soft X-ray emission that appears to fill holes in the column density distribution of gas-and-dust derived from submillimeter maps. • Our data shed new light on two quasi-symmetric lobes situated to Galactic north and south of Sgr A? . The Northern lobe shows a bright and sharp transition at its edge, suggesting the presence of a shock. Such features are possibly the remnant of the SN that generated SGR J1745-2900 or the PWN candidate G359.945-0.044. Alternatively, the lobes might constitute a long-lived bipolar structure produced by an isotropic outflow produced by either 1) the cumulative winds from the young stars of the central cluster, 2) a wind associated with the accretion flow onto Sgr A? , or 3) the same process that generated the X-ray reflection nebulae (if such activity has been recurrent over the past millennia). • The uniform X-ray colour of the superbubble G359.9-0.125, its sharp external edge and its being filled with S XV emitting plasma suggest that the soft X-ray features southwest of Sgr A? form a unique shell-like structure with total energy Eth ∼ 1051 erg, therefore making it a superbubble candidate in the GC (high absorption indicates that G359.9-0.125 is located at the GC). Alternatively, it might be the remnant of a very energetic event at the GC, such as a tidal disruption event. • We discover new evidence for the GC superbubble G0.1-0.1, also known as the arc-bubble from mid-IR observations: its soft Xray (e.g. S XV) emission completely fills the mid-IR bubble, and indicates a thermal energy as large as Eth ∼ 1.5 × 1051 erg. At present the Quintuplet cluster, which is moving at very high speed through the CMZ, is located at the border of the superbubble. However, it was more centrally located a few 104 yr ago and it could have, at least in part, energised it. We do not observe similar soft Xray emission trailing the Arches cluster, but this might be ascribed to its younger age. • We suggest that the Galactic Center Lobe might be a magnetic structure filled with warm, soft, X-ray-emitting plasma. In fact, we observe: i) enhanced soft X-ray emission at high Galactic latitudes; ii) enhanced soft X-ray emission at, and between, the longitudes of the Radio Arc and the Chimney associated with Sgr C, corresponding to the east and west foot-points of the GCL; iii) a sharp edge (at l = 359.63◦ , b = 0.06◦ and l = 359.55◦ , b = 0.46◦ ), running parallel to the nonthermal ripple filament (G359.54+0.18) and Sgr C thread, defining the western border of the enhanced soft X-ray emission. The GCL could be the relatively small base of an even larger structure, the so-called Fermi Bubbles. Additional observations will be needed to clarify this. • A new very faint X-ray transient, XMMU J17450.3-291445, has been discovered during the new XMM-Newton campaign to reach a peak luminosity of LX ∼ 1035 erg s−1 for ∼ 2 hr (Soldi et al. 2014). ACKNOWLEDGMENTS This research has made use both of data obtained with XMMNewton, an ESA science mission with instruments and contributions directly funded by ESA Member States and NASA, and data obtained from the Chandra Data Archive. We kindly acknowledge Sergio Molinari for providing the Herschel map, Casey Law for the GBT images and Namir Kassim for the VLA 90-cm map. GP acknowledges Roland Crocker, Barbara De Marco and Pierre Maggi for useful discussions. GP also acknowledges Frederick Baganoff and Nanda Rea for discussions about the origin of the lobes and the association with the SNR of SGR J1745-2900. We thank the referee for a careful reading of the paper. GP acknowledges support via an EU Marie Curie Intra-European Fellowship under contract no. FP7-PEOPLE-2012-IEF-331095. The GC XMM-Newton monitoring project is partially supported by the Bundesministerium für Wirtschaft und Technologie/Deutsches Zentrum für Luft- und Raumfahrt (BMWI/DLR, FKZ 50 OR 1408) and the Max Planck Society. Partial support through the COST action MP0905 Black Holes in a Violent Universe is acknowledged. The authors thank the ISSI in Bern. REFERENCES Aharonian, F., Akhperjanian, A. G., Aye, K.-M., et al. 2005, A&A, 432, L25 Alexander, T., & Hopman, C. 2003, ApJ, 590, L29 Amo-Baladrón, M. A., Martı́n-Pintado, J., Morris, M. R., Muno, M. P., & Rodrı́guez-Fernández, N. J. 2009, ApJ, 694, 943 Anantharamaiah, K. R., Pedlar, A., Ekers, R. D., & Goss, W. M. 1991, MNRAS, 249, 262 Ayal, S., Livio, M., & Piran, T. 2000, ApJ, 545, 772 Baganoff, F. K., Bautz, M. W., Brandt, W. N., et al. 2001, Nature, 413, 45 Baganoff, F. K., Maeda, Y., Morris, M., et al. 2003, ApJ, 591, 891 Bally, J., & Yusef-Zadeh, F. 1989, ApJ, 336, 173 Bamba, A., Yokogawa, J., Sakano, M., & Koyama, K. 2000, PASJ, 52, 259 Bamba, A., Murakami, H., Senda, A., et al. 2002, arXiv:astroph/0202010 Bamba, A., Yamazaki, R., Kohri, K., et al. 2009, ApJ, 691, 1854 Barrière, N. M., Tomsick, J. A., Baganoff, F. K., et al. 2014, ApJ, 786, 46 Bélanger, G., Goldwurm, A., Melia, F., et al. 2005, ApJ, 635, 1095 Bélanger, G., Goldwurm, A., Renaud, M., et al. 2006, ApJ, 636, 275 Blanton, M. C. 2008, Ph.D. Thesis, Bland-Hawthorn, J., & Cohen, M. 2003, ApJ, 582, 246 c 0000 RAS, MNRAS 000, 1–?? 35 EPIC-pn Exp Exp Mod ID EPIC-MOS1 Exp Exp Mod ID EPIC-MOS2 Exp Exp Mod ID EPIC-pn mode filter EPIC-MOS1 mode filter EPIC-MOS2 mode filter rev 0694640101 0694640201 0694640301 0694640401 0694640501 0694640601 0694640701 0694640801 0694640901 0694641001 0694641101 0694641201 0694641301 0694641401 0694641501 0694641601 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 2335 S S S S S S S S S S S S S S S S 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 S S S S S S S S S S S S S S S S NEW CMZ XMM-Newton scan 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF med med med med med med med med med med med med med med med med FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF med med med med med med med med med med med med med med med med FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF med med med med med med med med med med med med med med med med 6.0 6.0 6.0 8.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 2.0 2.0 2.0 2.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 0112970101 0112970201 0112970401 0112970501 0112970701 0112970801 0112971001 0112971301 0112971501 0112971601 0112971701 0112971801 0112971901 0112972101 0145 0145 0143 0144 0139 0144 0145 0143 0240 0240 0240 0240 0240 0318 U S S S S S S S S S S S S S 002 003 003 003 003 003 003 003 003 003 003 003 003 003 U S S S S S S S S S S S S S OLD CMZ XMM-Newton scan 002 U 002 FF 001 S 002 eFF 001 S 002 eFF 001 S 002 eFF 001 S 002 eFF 001 S 002 eFF 001 S 002 FF 001 S 408 SW 001 S 002 eFF 001 S 002 eFF 011 S 010 SW 001 S 002 eFF 001 S 002 eFF 001 S 002 eFF med med med med med med tck med med med med med med med FF FF FF FF FF FF FF TU FF FF TU FF FF FF med med med med med med med med med med med med med med FF FF FF FF FF FF FF RF FF FF SW FF FF FF med med med med med med med med med med med med med med 6.0 6.0 6.0 6.0 6.0 6.0 5.0 6.0 6.0 6.0 6.0 6.0 5.0 6.0 2.0 2.0 2.0 2.0 2.0 1.5 1.5 1.5 2.0 2.0 2.0 2.0 1.5 1.5 2.0 2.0 2.0 2.0 2.0 1.5 1.5 1.5 2.0 2.0 2.0 2.0 1.5 1.5 FF FF tck tck FF FF med med FF FF med med 6.0 6.0 2.0 2.0 2.0 2.0 eF eF FF FF F med F med med med FF FF FF FF med med med med FF FF FF FF med med med med 6.0 6.0 6.0 6.0 2.0 2.0 2.0 1.5 2.0 2.0 2.0 1.5 FF FF med med FF FF med med FF FF med med 6.0 7.0 2.0 2.0 2.0 2.0 FF FF FF FF med med med med FF FF FF FF med med med med FF FF FF FF med med med med 8.0 8.0 8.0 8.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 FF FF FF thn thn med FF FF FF thn thn med FF FF FF thn thn med 7.0 7.0 8.0 2.0 2.5 2.0 2.0 2.5 2.0 FF FF FF med med med FF FF FF med med med FF FF FF med med med 8.0 8.0 8.0 2.5 2.5 2.5 2.5 2.5 2.5 FF FF FF FF FF FF FF med med med med med med med FF FF FF FF FF FF FF med med med med med med med FF FF FF FF FF FF FF med med med med med med med 7.0 7.0 8.0 7.0 7.0 8.0 7.0 2.0 2.0 2.0 2.0 2.0 2.5 1.8 2.0 2.0 2.0 2.0 2.0 2.5 1.8 FF FF FF FF FF med med med med med FF FF FF FF FF med med med med med FF FF FF FF FF med med med med med 7.0 7.0 6.0 7.0 6.0 2.0 2.0 1.8 1.8 2.0 2.0 2.0 1.8 1.8 2.0 0111350101 0111350301 0406 0516 U S 002 001 S S 0202670501 0202670601 0202670701 0202670801 0788 0789 0866 0867 U S S S 002 003 003 003 U S S S 0302882601 0302884001 1139 1236 S S 003 003 S S 0402430301 0402430401 0402430701 0504940201 1339 1340 1338 1418 S U S S 001 002 001 003 S U S S 0511000301 0511000401 0505670101 1508 1610 1518 S S U 003 003 002 S U U 0554750401 0554750501 0554750601 1705 1706 1707 S S U 003 003 002 S S S 0604300601 0604300701 0604300801 0604300901 0604301001 0658600101 0658600201 2069 2070 2071 2072 2073 2148 2148 S U U S S S S 003 002 002 003 003 001 001 S S U S S S S 0674600601 0674600701 0674600801 0674601001 0674601101 2245 2246 2248 2249 2247 S S S S S 003 003 003 003 003 S S S S U Pointing toward Sgr A? 2002 006 S 005 006 S 005 2004 003 U 003 001 S 002 001 S 002 001 S 002 2006 001 S 002 001 S 002 2007 002 S 003 002 U 002 002 S 003 001 S 002 2008 001 S 002 002 U 002 002 U 002 2009 001 S 002 001 S 002 001 S 002 2011 001 S 002 001 S 002 002 U 002 001 S 002 001 S 002 002 S 003 002 S 003 2012 001 S 002 001 S 002 001 S 002 001 S 002 002 U 002 pn c/s Threshold M1 M2 c/s c/s OBSID Table 8. List of all XMM-Newton observations considered in this work. Exposure Mode: U, S stand for unscheduled and scheduled, respectively. Filters: Med, thn, tck stand for medium, thin and thick filters, respectively. FF, eFF, SW, Ti, TU stand for full frame, extended full frame, small window, timing and time uncompressed, respectively. Borkowski, K. J., Reynolds, S. P., Hwang, U., et al. 2013, ApJ, 771, LL9 Breitschwerdt, D., McKenzie, J. F., & Voelk, H. J. 1991, A&A, 245, 79 Capelli, R., Warwick, R. S., Porquet, D., Gillessen, S., & Predehl, P. 2011, A&A, 530, AA38 Capelli, R., Warwick, R. S., Porquet, D., Gillessen, S., & Predehl, P. 2012, A&A, 545, A35 Cash, W. 1979, ApJ, 228, 939 Castor, J., McCray, R., & Weaver, R. 1975, ApJ, 200, L107 Castro, M., Maiolino, T., D’Amico, F., et al. 2013, arXiv:1302.6213 Chuss, D. T., Davidson, J. A., Dotson, J. L., et al. 2003, ApJ, 599, c 0000 RAS, MNRAS 000, 1–?? 1116 Clark, P. C., Glover, S. C. O., Ragan, S. E., Shetty, R., & Klessen, R. S. 2013, ApJ, 768, L34 Clavel, M., Terrier, R., Goldwurm, A., et al. 2013, A&A, 558, A32 Clavel, M., Soldi, S., Terrier, R., et al. 2014, MNRAS, 443, L129 Coil, A. L., & Ho, P. T. P. 2000, ApJ, 533, 245 Cotera, A. S., Erickson, E. F., Colgan, S. W. J., et al. 1996, ApJ, 461, 750 Coti Zelati, F., Rea, N., Papitto, A., et al. 2015, MNRAS, 449, 2685 Crocker, R. M., Jones, D. I., Melia, F., Ott, J., & Protheroe, R. J. 2010, Nature, 463, 65 Crocker, R. M., & Aharonian, F. 2011, Physical Review Letters, 36 G. Ponti et al. OBSID rev 0030540101 0144220101 0152920101 0144630101 0203930101 0205240101 0304220301 0304220101 0303210201 0302882501 0302882701 0302882801 0302882901 0302883001 0302883101 0302883201 0305830701 0302883901 0302884101 0302884201 0302884301 0302884401 0302884501 0406580201 0410580401 0410580501 0400340101 0506291201 0504940101 0504940401 0504940501 0504940601 0504940701 0511010701 0511000101 0511000501 0511000701 0511000901 0511001101 0511001301 0511000201 0511000601 0511000801 0511001001 0511001201 0511001401 0505870301 0603850201 0655670101 0504 0596 0607 0688 0868 0956 1048 1063 1065 1139 1139 1139 1139 1139 1139 1139 1157 1236 1236 1236 1236 1236 1236 1241 1243 1245 1244 1322 1418 1418 1418 1418 1418 1505 1508 1508 1508 1508 1508 1508 1510 1510 1512 1512 1512 1512 1511 1891 2065 EPIC-pn Exp Exp Mod ID S U S S S S S S S S S S S S S S S S S S S S S S N N S N S S S S S S S S S S S S S S S S S S S S N 003 002 003 003 003 003 004 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 000 000 003 000 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 003 000 EPIC-MOS1 Exp Exp Mod ID S U S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S EPIC-MOS2 Exp Exp Mod ID EPIC-pn mode filter Other observations of the CMZ 001 S 002 SW 002 U 002 SW 001 S 002 FF 001 S 002 SW 001 S 002 eF 001 S 002 FF 002 S 003 SW 001 S 002 SW 001 S 002 SW 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 002 S 003 Ti 002 S 003 Ti 001 S 002 FF 001 S 002 Ti 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 FF 001 S 002 Ti tck med tck med F med med med med med med med med med med med med med med med med med med med med tck tck med med med med med med med med thn thn thn thn thn thn thn thn thn thn thn thn med med med EPIC-MOS1 mode filter SW FF FF TU FF FF FF FF TU FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF tck med tck med med med med med med med med med med med med med med med med med med med med med tck tck med med med med med med med med thn thn thn thn thn thn thn thn thn thn thn thn med med med EPIC-MOS2 mode filter SW FF FF SW FF FF FF FF TU FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF TU TU FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF tck med tck med med med med med med med med med med med med med med med med med med med med med tck tck med med med med med med med med thn thn thn thn thn thn thn thn thn thn thn thn med med med pn c/s Threshold M1 M2 c/s c/s 6.0 6.0 6.0 2.0 1.5 1.5 2.0 1.5 1.5 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 7.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 8.0 6.0 6.0 6.0 6.0 1.5 1.5 1.5 1.5 1.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 1.5 2.0 2.0 2.0 2.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.5 2.0 2.0 2.0 2.5 2.5 2.0 2.5 2.5 2.5 2.5 2.5 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.5 1.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 1.5 2.0 2.0 2.0 2.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.5 2.0 2.0 2.0 2.5 2.5 2.0 2.5 2.5 2.5 2.5 2.5 2.5 2.0 2.0 2.0 Table 9. List of all XMM-Newton observations considered in this work. Exposure Mode: U, S stand for unscheduled and scheduled, respectively. Filters: Med, thn, tck stand for medium, thin and thick filters, respectively. FF, eFF, SW, Ti, TU stand for full frame, extended full frame, small window, timing and time uncompressed, respectively. 106, 101102 Crocker, R. M. 2012, MNRAS, 423, 3512 Decourchelle, A. 2003, XMM-Newton Proposal, 210 Degenaar, N., & Wijnands, R. 2010, A&A, 524, A69 Degenaar, N., Wijnands, R., Cackett, E. M., et al. 2012, A&A, 545, A49 Degenaar, N., Reynolds, M. T., Miller, J. M., Kennea, J. A., & Wijnands, R. 2013, The Astronomer’s Telegram, 5006, 1 Del Santo, M., Sidoli, L., Bazzano, A., et al. 2006, A&A, 456, 1105 Do, T., Lu, J. R., Ghez, A. M., et al. 2013, ApJ, 764, 154 Dong, H., Wang, Q. D., Cotera, A., et al. 2011, MNRAS, 417, 114 Dong, H., Wang, Q. D., & Morris, M. R. 2012, MNRAS, 425, 884 Downes, D., & Maxwell, A. 1966, ApJ, 146, 653 Dubner, G., Giacani, E., & Decourchelle, A. 2008, A&A, 487, 1033 Dutra, C. M., & Bica, E. 2000, A&A, 359, L9 Dutra, C. M., Ortolani, S., Bica, E., et al. 2003, A&A, 408, 127 Dwelly, T., & Ponti, G. 2013, TheAstronomer’s Telegram, 5008, 1 Ebisawa, K., Maeda, Y., Kaneda, H., & Yamauchi, S. 2001, Science, 293, 1633 Ferreira, S. E. S., & de Jager, O. C. 2008, A&A, 478, 17 Ferrière, K. 2009, A&A, 505, 1183 Ferrière, K. 2011, The Galactic Center: a Window to the Nuclear Environment of Disk Galaxies, 439, 39 Figer, D. F., Kim, S. S., Morris, M., et al. 1999, ApJ, 525, 750 Freyberg, M. J., Briel, U. G., Dennerl, K., et al. 2004, Proc. SPIE, 5165, 112 Fukuoka, R., Koyama, K., Ryu, S. G., & Tsuru, T. G. 2009, PASJ, 61, 593 Gaensler, B. M., Pivovaroff, M. J., & Garmire, G. P. 2001, ApJ, 556, L107 Gaensler, B. M., van der Swaluw, E., Camilo, F., et al. 2004, ApJ, 616, 383 Geballe, T. R., Wade, R., Krisciunas, K., Gatley, I., & Bird, M. C. 1987, ApJ, 320, 562 Genzel, R., Eisenhauer, F., & Gillessen, S. 2010, Reviews of Modern Physics, 82, 3121 Goldwurm, A., Brion, E., Goldoni, P., et al. 2003, ApJ, 584, 751 Gray, A. D. 1994, MNRAS, 270, 835 Green, D. A. 2014, arXiv:1409.0637 Haberl, F., & Pietsch, W. 1999, A&AS, 139, 277 Haberl, F., Sturm, R., Ballet, J., et al. 2012, A&A, 545, A128 Habibi, M., Stolte, A., Brandner, W., Hußmann, B., & Motohara, K. 2013, A&A, 556, AA26 Habibi, M., Stolte, A., & Harfst, S. 2014, A&A, 566, AA6 Hales, C. A., Gaensler, B. M., Chatterjee, S., van der Swaluw, E., & Camilo, F. 2009, ApJ, 706, 1316 Harrison, F. A., Craig, W. W., Christensen, F. E., et al. 2013, ApJ, c 0000 RAS, MNRAS 000, 1–?? 37 OBSID obs date 0694640101 0694640201 0694640301 0694640401 0694640501 0694640601 0694640701 0694640801 0694640901 0694641001 0694641101 0694641201 0694641301 0694641401 0694641501 0694641601 2012-09-07 2012-08-30 2012-08-31 2012-09-02 2012-09-05 2012-09-06 2012-10-02 2012-10-06 2012-09-12 2012-09-23 2012-09-24 2012-09-26 2012-09-26 2012-09-30 2012-10-06 2012-10-08 0112970101 0112970201 0112970401 0112970501 0112970701 0112970801 0112971001 0112971301 0112971501 0112971601 0112971701 0112971801 0112971901 0112972101 2000-09-23 2000-09-23 2000-09-19 2000-09-21 2000-09-11 2000-09-21 2000-09-24 2000-09-19 2001-04-01 2001-03-31 2001-03-31 2001-04-01 2001-04-01 2001-09-04 0111350101 0111350301 2002-02-26 2002-10-03 0202670501 0202670501 0202670501 0202670601 0202670601 0202670701 0202670801 2004-03-28 2004-03-28 2004-03-30 2004-03-30 2004-03-30 2004-08-31 2004-09-02 0302882601 0302884001 2006-02-27 2006-09-08 0402430301 0402430401 0402430701 0504940201 2007-04-01 2007-04-03 2007-03-30 2007-09-06 0511000301 0511000401 0505670101 2008-03-03 2008-09-23 2008-03-23 0554750401 0554750501 0554750601 2009-04-01 2009-04-03 2009-04-05 0604300601 0604300701 0604300801 0604300901 0604301001 0658600101 0658600201 2011-03-28 2011-03-30 2011-04-01 2011-04-03 2011-04-05 2011-08-31 2011-09-01 0674600601 0674600701 0674600801 0674601001 0674601101 2012-03-13 2012-03-15 2012-03-19 2012-03-21 2012-03-17 Exp pn Exp M1 Exp M2 Exp pn Exp M1 Exp M2 NEW CMZ XMM-Newton scan 41978 43452 43605 45035 46616 46619 40041 41616 41619 52954 51442 51460 44976 46606 46621 40042 41614 41621 42539 44099 44120 40041 41616 41619 43031 44617 44604 46021 47607 47620 40041 41616 41619 40008 41559 41577 53842 56260 56348 45816 46751 46920 49746 51483 51486 40005 41585 41585 38739 45038 40041 38736 32935 40042 42539 40041 42202 46041 40041 40008 46667 32466 39167 27250 38980 46616 41616 40073 33180 41614 44117 41616 43784 47614 41616 41588 48012 33767 40518 27795 38985 46619 41619 40075 33185 41621 44120 41619 43786 47620 41619 41598 48018 33770 40507 27803 OLD CMZ XMM-Newton scan 12870 15806 15611 13499 17394 17392 25411 29365 29391 21119 24914 24911 19518 23419 23413 19969 23892 23892 12599 16492 16482 12800 0 13091 20293 25020 25017 0 3996 3949 11000 0 11799 9927 14513 14542 4698 9191 9191 21687 26039 26055 12252 12999 21880 10289 19383 13462 8774 0 6752 0 0 1900 4147 20130 14679 16894 23849 14084 23221 17198 12529 0 7017 0 0 2069 8379 23515 2 14637 16892 23847 14081 23218 17198 12529 0 7017 0 0 2069 8379 3517 40030 8261 52118 9877 52120 9880 45847 0 0 56926 0 78857 91795 0 0 0 0 0 78921 93131 0 0 0 0 0 78915 93126 1700 4787 3160 6365 3163 6370 50962 36886 21240 7392 50955 36892 22820 8949 50958 36876 22825 8960 3305 5058 64200 4863 4358 65143 4868 4342 65153 31934 38634 31485 33358 40216 37464 33363 40218 37466 28768 32872 33771 19941 32571 47653 39634 30121 39149 36149 21140 33917 49169 41109 30119 39156 36129 21143 33914 49177 41115 8594 6802 16784 19841 8956 9296 8209 18358 21416 8173 9301 8212 18358 21419 8178 Pointing toward Sgr A? 52105 52120 16960 16996 2004 110170 5733 6087 0 107784 108572 0 650 848 112204 585 538 0 120863 122251 127470 132469 132503 130951 132997 133036 2006 4937 6563 6568 4987 6563 6570 2007 101319 93947 94022 93594 97566 96461 32338 33912 33917 11092 12649 12652 2008 5057 6615 6620 5058 4358 4342 96601 97787 97787 2009 38034 39614 39619 42434 44016 44018 32837 38816 38818 2011 45306 48467 48491 42305 48579 48584 37321 38642 38494 36568 37589 37573 48210 47757 47646 47585 49169 49159 51324 52903 52908 2012 19594 21167 21172 14040 15616 15618 21041 22615 22618 22034 23616 23619 25682 24638 24628 40030 15377 Table 10. List of all XMM-Newton observations considered in this work. Total and cleaned exposure time (in seconds) for each camera, respectively. 770, 103 Heard, V., & Warwick, R. S. 2013a, MNRAS, 428, 3462 Heard, V., & Warwick, R. S. 2013b, MNRAS, 434, 1339 Heinke, C. O., Tomsick, J. A., Yusef-Zadeh, F., & Grindlay, J. E. 2009, ApJ, 701, 1627 Henze, M., Pietsch, W., Haberl, F., et al. 2014, A&A, 563, A2 Hewitt, J. W., Yusef-Zadeh, F., & Wardle, M. 2008, ApJ, 683, 189 Ho, P. T. P., Jackson, J. M., Barrett, A. H., & Armstrong, J. T. 1985, ApJ, 288, 575 Inui, T., Koyama, K., Matsumoto, H., & Tsuru, T. G. 2009, PASJ, 61, 241 Johnson, J. L., Greif, T. H., Bromm, V., Klessen, R. S., & Ippolito, J. 2009, MNRAS, 399, 37 Kaneda, H., Makishima, K., Yamauchi, S., et al. 1997, ApJ, 491, 638 c 0000 RAS, MNRAS 000, 1–?? Kaspi, V. M., Archibald, R. F., Bhalerao, V., et al. 2014, ApJ, 786, 84 Kassim, N. E., & Frail, D. A. 1996, MNRAS, 283, L51 Kavanagh, P. J., Sasaki, M., & Points, S. D. 2012, A&A, 547, A19 Khokhlov, A., & Melia, F. 1996, ApJ, 457, L61 Koch, E. W., Bahramian, A., Heinke, C. O., et al. 2014, MNRAS, 442, 372 Koyama, K., Hyodo, Y., Inui, T., et al. 2007, PASJ, 59, 245 Koyama, K., Inui, T., Hyodo, Y., et al. 2007, PASJ, 59, 221 Koyama, K., Maeda, Y., Sonobe, T., et al. 1996, PASJ, 48, 249 Koyama, K., Hyodo, Y., Inui, T., et al. 2007, PASJ, 59, 245 Koyama, K., Inui, T., Matsumoto, H., & Tsuru, T. G. 2008, PASJ, 60, 201 Koyama, K., Takikawa, Y., Hyodo, Y., et al. 2009, PASJ, 61, 255 Krivonos, R. A., Tomsick, J. A., Bauer, F. E., et al. 2014, ApJ, 38 G. Ponti et al. OBSID obs date 0030540101 0144220101 0152920101 0144630101 0203930101 0205240101 0304220301 0304220101 0303210201 0302882501 0302882701 0302882801 0302882901 0302883001 0302883101 0302883201 0305830701 0302883901 0302884101 0302884201 0302884301 0302884401 0302884501 0406580201 0410580401 0410580501 0400340101 0506291201 0504940101 0504940401 0504940501 0504940601 0504940701 0511010701 0511000101 0511000501 0511000701 0511000901 0511001101 0511001301 0511000201 0511000601 0511000801 0511001001 0511001201 0511001401 0505870301 0603850201 0655670101 2002-09-09 2003-03-12 2003-04-02 2003-09-11 2004-09-04 2005-02-26 2005-08-29 2005-09-29 2005-10-02 2006-02-27 2006-02-27 2006-02-27 2006-02-27 2006-02-27 2006-02-27 2006-03-29 2006-04-04 2006-09-08 2006-09-08 2006-09-08 2006-09-09 2006-09-09 2006-09-09 2006-09-18 2006-09-22 2006-09-26 2006-09-24 2007-02-27 2007-09-06 2007-09-06 2007-09-06 2007-09-06 2007-09-06 2008-02-27 2008-03-03 2008-03-04 2008-03-04 2008-03-04 2008-03-04 2008-03-04 2008-09-23 2008-09-23 2008-09-27 2008-09-27 2008-09-27 2008-09-27 2008-03-10 2010-04-07 2011-03-19 Exp pn Exp M1 Exp M2 Other observations of the CMZ 27689 27842 27844 46746 49905 49843 50182 51639 51774 8469 0 8661 46544 50438 50446 46919 50625 50604 20031 20213 20226 8051 8237 8250 23472 0 0 7561 9176 9178 5237 6851 6869 5937 7558 7571 5936 7566 7569 5937 7540 7558 9814 11432 11448 4896 6518 6526 6399 11266 11256 4987 6565 6568 4987 6565 6570 4987 6565 6570 4987 6565 6568 4036 5616 5621 6787 8364 8370 28034 29607 29609 0 32558 0 0 32116 0 40001 41575 41580 0 38616 38621 5058 6615 6620 5058 6615 6620 5057 6615 6620 5058 6615 6620 5058 6615 6620 7455 9004 9004 6943 8500 8500 5058 6615 6620 5058 6615 6620 5058 6615 6620 5057 6615 6620 5058 6615 6620 5058 6615 6620 5058 6615 6620 5035 6602 6620 5034 6615 6620 5034 6615 6620 5034 6615 6620 29885 31614 31494 22503 21643 21663 0 103934 103954 Exp pn Exp M1 Exp M2 27339 28820 48486 0 39078 14946 6417 5621 23 6364 2937 5537 4437 3137 8614 3898 0 4787 4000 4987 4987 4037 6787 13896 0 0 16312 0 4958 5058 5006 1720 4558 5803 546 4658 4506 5058 5057 3800 5058 5058 5035 5034 5034 5034 7249 16919 0 27495 31550 50082 316 43003 15251 6615 5816 314 7999 4564 7164 6066 4766 10248 5547 1028 6365 5578 6565 6565 5616 8364 14809 32367 30108 17481 30937 6515 6615 6563 3175 6115 7362 796 6215 6063 6514 6615 5132 6615 6615 6615 6615 6615 6615 7250 18271 80716 27494 31461 50097 311 43013 15243 6620 5821 315 8002 4569 7171 6069 4771 10258 5539 1028 6368 5583 6570 6568 5621 8369 14814 32326 30096 17486 30937 6520 6620 6568 3181 6120 7368 800 6220 6068 6518 6620 5137 6620 6620 6620 6620 6620 6620 7255 18266 80729 Table 11. List of all XMM-Newton observations considered in this work. Total and cleaned exposure time (in seconds) for each camera, respectively. 781, 107 Kroupa, P. 2002, Science, 295, 82 Kuntz, K. D., & Snowden, S. L. 2008, A&A, 478, 575 Kuulkers, E., Shaw, S. E., Paizis, A., et al. 2007, A&A, 466, 595 Lakićević, M., van Loon, J. T., Meixner, M., et al. 2014, arXiv:1410.5709 Lang, C. C., Morris, M., & Echevarria, L. 1999, ApJ, 526, 727 Lang, C. C., Goss, W. M., & Morris, M. 2002, AJ, 124, 2677 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, AJ, 119, 207 Launhardt, R., Zylka, R., & Mezger, P. G. 2002, A&A, 384, 112 Law, C., & Yusef-Zadeh, F. 2004, ApJ, 611, 858 Law, C. J., Backer, D., Yusef-Zadeh, F., & Maddalena, R. 2009, ApJ, 695, 1070 Law, C. J., Brentjens, M. A., & Novak, G. 2011, ApJ, 731, 36 Levine, D., Morris, M., & Figer, D. 1999, The Universe as Seen by ISO, 427, 699 Lu, F. J., Wang, Q. D., & Lang, C. C. 2003, AJ, 126, 319 Lu, F. J., Yuan, T. T., & Lou, Y.-Q. 2008, ApJ, 673, 915 Lu, J. R., Do, T., Ghez, A. M., et al. 2013, ApJ, 764, 155 Maeda, Y., Baganoff, F. K., Feigelson, E. D., et al. 2002, ApJ, 570, 671 Markoff, S. 2010, Proceedings of the National Academy of Science, 107, 7196 Marquez-Lugo, R. A., & Phillips, J. P. 2010, MNRAS, 407, 94 Mauerhan, J. C., Muno, M. P., Morris, M. R., Stolovy, S. R., & Cotera, A. 2010, ApJ, 710, 706 Mac Low, M.-M., & McCray, R. 1988, ApJ, 324, 776 Mereghetti, S., Sidoli, L., & Israel, G. L. 1998, A&A, 331, L77 Misanovic, Z., Pietsch, W., Haberl, F., et al. 2006, A&A, 448, 1247 Molaro, M., Khatri, R., & Sunyaev, R. A. 2014, A&A, 564, AA107 Molinari, S., Bally, J., Noriega-Crespo, A., et al. 2011, ApJ, 735, L33 Mori, H., Maeda, Y., Pavlov, G. G., Sakano, M., & Tsuboi, Y. 2005, Advances in Space Research, 35, 1137 Mori, H., Tsuru, T. G., Hyodo, Y., Koyama, K., & Senda, A. 2008, PASJ, 60, 183 Mori, H., Hyodo, Y., Tsuru, T. G., Nobukawa, M., & Koyama, K. 2009, PASJ, 61, 687 Mori, K., Gotthelf, E. V., Zhang, S., et al. 2013, ApJ, 770, LL23 Morris, M., & Yusef-Zadeh, F. 1985, AJ, 90, 2511 Morris, M., & Yusef-Zadeh, F. 1989, ApJ, 343, 703 Morris, M. 1990, Galactic and Intergalactic Magnetic Fields, 140, 361 Morris, M., & Serabyn, E. 1996, ARA&A, 34, 645 Morris, M., Baganoff, F., Muno, M., et al. 2003, Astronomische Nachrichten Supplement, 324, 167 Morris, M., Howard, C., Muno, M., et al. 2004, The Dense Interstellar Medium in Galaxies, 281 Morris, M. R., Meyer, L., & Ghez, A. M. 2012, Research in Astronomy and Astrophysics, 12, 995 Morris, M. R., Zhao, J.-H., & Goss, W. M. 2014a, IAU Symposium, 303, 369 Morris, M. R. 2014b, arXiv:1406.7859 c 0000 RAS, MNRAS 000, 1–?? 39 Mossoux, E., Grosso, N., Vincent, F. H., & Porquet, D. 2015, A&A, 573, A46 Muno, M. P., Baganoff, F. K., Bautz, M. W., et al. 2003, ApJ, 589, 225 Muno, M. P., Baganoff, F. K., Bautz, M. W., et al. 2004, ApJ, 613, 326 Muno, M. P., Pfahl, E., Baganoff, F. K., et al. 2005a, ApJ, 622, L113 Muno, M. P., Lu, J. R., Baganoff, F. K., et al. 2005b, ApJ, 633, 228 Muno, M. P., Bauer, F. E., Bandyopadhyay, R. M., & Wang, Q. D. 2006, ApJS, 165, 173 Muno, M. P., Baganoff, F. K., Brandt, W. N., Park, S., & Morris, M. R. 2007, ApJ, 656, L69 Muno, M. P., Baganoff, F. K., Brandt, W. N., Morris, M. R., & Starck, J.-L. 2008, ApJ, 673, 251 Muno, M. P., Bauer, F. E., Baganoff, F. K., et al. 2009, ApJS, 181, 110 Nakashima, S., Nobukawa, M., Tsuru, T. G., Koyama, K., & Uchiyama, H. 2010, PASJ, 62, 971 Natalucci, L., Bazzano, A., Cocchi, M., et al. 2004, A&A, 416, 699 Natalucci, L., Tomsick, J. A., Bazzano, A., et al. 2014, ApJ, 780, 63 Najarro, F., Krabbe, A., Genzel, R., et al. 1997, A&A, 325, 700 Neilsen, J., Nowak, M. A., Gammie, C., et al. 2013, ApJ, 774, 42 Nishiyama, S., Tamura, M., Hatano, H., et al. 2009, ApJ, 690, 1648 Nishiyama, S., Yasui, K., Nagata, T., et al. 2013, ApJ, 769, L28 Nobukawa, M., Tsuru, T. G., Takikawa, Y., et al. 2008, PASJ, 60, 191 Nobukawa, M., Sawada, M., Matsumoto, H., Tsuru, T. G., & Koyama, K. 2009, Advances in Space Research, 43, 1045 Nobukawa, M., Koyama, K., Tsuru, T. G., Ryu, S. G., & Tatischeff, V. 2010, PASJ, 62, 423 Nobukawa, M., Ryu, S. G., Tsuru, T. G., & Koyama, K. 2011, ApJ, 739, L52 Nord, M. E., Lazio, T. J. W., Kassim, N. E., et al. 2004, AJ, 128, 1646 Novak, G., Chuss, D. T., Renbarger, T., et al. 2003, ApJ, 583, L83 Nynka, M., Hailey, C. J., Reynolds, S. P., et al. 2014, ApJ, 789, 72 Ohnishi, T., Koyama, K., Tsuru, T. G., et al. 2011, PASJ, 63, 527 Ostriker, J. P., & McKee, C. F. 1988, Reviews of Modern Physics, 60, 1 Paizis, A., Ebisawa, K., Takahashi, H., et al. 2009, PASJ, 61, 107 Park, S., Muno, M. P., Baganoff, F. K., et al. 2004, ApJ, 603, 548 Park, S., Muno, M. P., Baganoff, F. K., et al. 2005, ApJ, 631, 964 Paumard, T., Maillard, J. P., Morris, M., & Rigaut, F. 2001, A&A, 366, 466 Pavlinsky, M. N., Grebenev, S. A., & Sunyaev, R. A. 1994, ApJ, 425, 110 Pedlar, A., Anantharamaiah, K. R., Ekers, R. D., et al. 1989, ApJ, 342, 769 Piraino, S., Santangelo, A., Kaaret, P., et al. 2012, A&A, 542, L27 Phillips, J. P., & Marquez-Lugo, R. A. 2010, MNRAS, 409, 701 Ponti, G., Terrier, R., Goldwurm, A., Belanger, G., & Trap, G. 2010, ApJ, 714, 732 Ponti, G., Fender, R. P., Begelman, M. C., et al. 2012, MNRAS, 422, L11 Ponti, G., Morris, M. R., Terrier, R., & Goldwurm, A. 2013, Cosmic Rays in Star-Forming Environments, 34, 331 Ponti, G., Morris, M. R., Clavel, M., et al. 2014, IAU Symposium, c 0000 RAS, MNRAS 000, 1–?? 303, 333 Ponti, G., Bianchi, S., Muñoz-Darias, T., et al. 2015, MNRAS, 446, 1536 Porquet, D., Rodriguez, J., Corbel, S., et al. 2003a, A&A, 406, 299 Porquet, D., Decourchelle, A., & Warwick, R. S. 2003b, A&A, 401, 197 Porquet, D., Grosso, N., Burwitz, V., et al. 2005a, A&A, 430, L9 Porquet, D., Grosso, N., Bélanger, G., et al. 2005, A&A, 443, 571 Porquet, D., Grosso, N., Predehl, P., et al. 2008, A&A, 488, 549 Predehl, P., & Kulkarni, S. R. 1995, A&A, 294, L29 Price, S. D., Egan, M. P., Carey, S. J., Mizuno, D. R., & Kuchar, T. A. 2001, AJ, 121, 2819 Pierce-Price, D., Richer, J. S., Greaves, J. S., et al. 2000, ApJ, 545, L121 Quataert, E., & Loeb, A. 2005, ApJ, 635, L45 Rea, N., Esposito, P., Pons, J. A., et al. 2013, ApJ, 775, L34 Reich, W., & Fuerst, E. 1984, A&AS, 57, 165 Reid, M. J. 1993, ARA&A, 31, 345 Reid, M. J., Menten, K. M., Zheng, X. W., Brunthaler, A., & Xu, Y. 2009, ApJ, 705, 1548 Renaud, M., Paron, S., Terrier, R., et al. 2006, ApJ, 638, 220 Revnivtsev, M. G., Churazov, E. M., Sazonov, S. Y., et al. 2004, A&A, 425, L49 Revnivtsev, M., Sazonov, S., Gilfanov, M., Churazov, E., & Sunyaev, R. 2006, A&A, 452, 169 Revnivtsev, M., Sazonov, S., Churazov, E., et al. 2009, Nature, 458, 1142 Reynolds, M. T., & Miller, J. M. 2010, ApJ, 716, 1431 Rodrı́guez-Fernández, N. J., Martı́n-Pintado, J., & de Vicente, P. 2001, A&A, 377, 631 Roy, S., & Bhatnagar, S. 2006, Journal of Physics Conference Series, 54, 152 Sakano, M., Koyama, K., Murakami, H., Maeda, Y., & Yamauchi, S. 2002, ApJS, 138, 19 Sakano, M., Warwick, R. S., Decourchelle, A., & Predehl, P. 2003a, MNRAS, 340, 747 Sakano, M., Warwick, R. S., & Decourchelle, A. 2003b, Workshop on Galaxies and Clusters of Galaxies, 9 Sakano, M., Warwick, R. S., Hands, A., & Decourchelle, A. 2004, Memorie della Societa Astronomica Italiana, 75, 498 Sakano, M., Warwick, R. S., Decourchelle, A., & Wang, Q. D. 2005, MNRAS, 357, 1211 Sasaki, M., Breitschwerdt, D.,Baumgartner, V., & Haberl, F. 2011, A&A, 528, A136 Sawada, M., Tsujimoto, M., Koyama, K., et al. 2009, PASJ, 61, 209 Seiradakis, J. H., Lasenby, A. N., Yusef-Zadeh, F., Wielebinski, R., & Klein, U. 1985, Nature, 317, 697 Senda, A., Murakami, H., & Koyama, K. 2002, ApJ, 565, 1017 Senda, A., Murakami, H., & Koyama, K. 2003, Astronomische Nachrichten Supplement, 324, 151 Shibata, K., Tajima, T., Steinolfson, R. S., & Matsumoto, R. 1989, ApJ, 345, 584 Simpson, J. P., Colgan, S. W. J., Cotera, A. S., et al. 2007, ApJ, 670, 1115 Skinner, G. K., Foster, A. J., Willmore, A. P., & Eyles, C. J. 1990, MNRAS, 243, 72 Sofue, Y., & Handa, T. 1984, Nature, 310, 568 Sofue, Y. 1985, PASJ, 37, 697 Soldi, S., Walter, R., Eckert, D., et al. 2006, The Astronomer’s Telegram, 885, 1 40 G. Ponti et al. Soldi, S., Clavel, M., Goldwurm, A., et al. 2014, IAU Symposium, 303, 126 Staguhn, J., Stutzki, J., Uchida, K. I., & Yusef-Zadeh, F. 1998, A&A, 336, 290 Stiele, H., Pietsch, W., Haberl, F., et al. 2011, A&A, 534, A55 Stolte, A., Hußmann, B., Morris, M. R., et al. 2014, ApJ, 789, 115 Sturm, R., Haberl, F., Pietsch, W., et al. 2013, A&A, 558, A3 Su, M., Slatyer, T. R., & Finkbeiner, D. P. 2010, ApJ, 724, 1044 Sunyaev, R. A., Markevitch, M., & Pavlinsky, M. 1993, ApJ, 407, 606 Tanaka, Y., Koyama, K., Maeda, Y., & Sonobe, T. 2000, PASJ, 52, L25 Tanaka, Y. 2002, A&A, 382, 1052 Tanaka, K., Kamegai, K., Nagai, M., & Oka, T. 2007, PASJ, 59, 323 Tanaka, K., Oka, T., Nagai, M., & Kamegai, K. 2009, PASJ, 61, 461 Tang, S., & Wang, Q. D. 2005, ApJ, 628, 205 Tatischeff, V., Decourchelle, A., & Maurin, G. 2012, A&A, 546, AA88 Terrier, R., Ponti, G., Bélanger, G., et al. 2010, ApJ, 719, 143 Trap, G., Falanga, M., Goldwurm, A., et al. 2009, A&A, 504, 501 Trap, G., Goldwurm, A., Dodds-Eden, K., et al. 2011, A&A, 528, A140 Tsuboi, M., Inoue, M., Handa, T., et al. 1986, AJ, 92, 818 Tsuru, T. G., Nobukawa, M., Nakajima, H., et al. 2009, PASJ, 61, 219 Tüllmann, R., Gaetz, T. J., Plucinsky, P. P., et al. 2011, ApJS, 193, 31 Uchida, Y., Sofue, Y., & Shibata, K. 1985, Nature, 317, 699 Uchida, K., Morris, M., & Serabyn, E. 1990, ApJ, 351, 443 Uchida, K., Morris, M., & Yusef-Zadeh, F. 1992, AJ, 104, 1533 Uchida, K. I., Morris, M. R., Serabyn, E., & Bally, J. 1994, ApJ, 421, 505 Uchiyama, H., Nobukawa, M., Tsuru, T., Koyama, K., & Matsumoto, H. 2011, PASJ, 63, 903 Uchiyama, H., Nobukawa, M., Tsuru, T. G., & Koyama, K. 2013, PASJ, 65, 19 Wang, Q. D., Lu, F., & Lang, C. C. 2002, ApJ, 581, 1148 Wang, D. X., Xiao, K., & Lei, W. H. 2002, MNRAS, 335, 655 Wang, Q. D., Gotthelf, E. V., & Lang, C. C. 2002, Nature, 415, 148 Wang, W., Jiang, Z. J., & Cheng, K. S. 2005, MNRAS, 358, 263 Wang, Q. D., Dong, H., & Lang, C. 2006a, MNRAS, 371, 38 Wang, Q. D., Lu, F. J., & Gotthelf, E. V. 2006b, MNRAS, 367, 937 Wang, Q. D., Dong, H., Cotera, A., et al. 2010, MNRAS, 402, 895 Wang, Q. D., Nowak, M. A., Markoff, S. B., et al. 2013, Science, 341, 981 Weaver, R., McCray, R., Castor, J., Shapiro, P., & Moore, R. 1977, ApJ, 218, 377 Werner, N., in’t Zand, J. J. M., Natalucci, L., et al. 2004, A&A, 416, 311 Wijnands, R., Miller, J. M., & Wang, Q. D. 2002, ApJ, 579, 422 Wijnands, R., in’t Zand, J. J. M., Rupen, M., et al. 2006, A&A, 449, 1117 Yamauchi, S., Shimizu, M., Nakashima, S., et al. 2014, arXiv:1409.4520 Yelda, S., Ghez, A. M., Lu, J. R., et al. 2014, ApJ, 783, 131 Yoast-Hull, T. M., Gallagher, J. S., III, & Zweibel, E. G. 2014, ApJ, 790, 86 Yusef-Zadeh, F., Morris, M., & Chance, D. 1984, Nature, 310, 557 Yusef-Zadeh, F., & Morris, M. 1987a, ApJ, 320, 545 Yusef-Zadeh, F., & Morris, M. 1987b, AJ, 94, 1178 Yusef-Zadeh, F., & Morris, M. 1987c, ApJ, 322, 721 Yusef-Zadeh, F., & Morris, M. 1988, ApJ, 329, 729 Yusef-Zadeh, F., Wardle, M., & Parastaran, P. 1997, ApJ, 475, L119 Yusef-Zadeh, F., Law, C., Wardle, M., et al. 2002, ApJ, 570, 665 Yusef-Zadeh, F., Hewitt, J. W., & Cotton, W. 2004, ApJS, 155, 421 Yusef-Zadeh, F., Wardle, M., Muno, M., Law, C., & Pound, M. 2005, Advances in Space Research, 35, 1074 Yusef-Zadeh, F., Muno, M., Wardle, M., & Lis, D. C. 2007, ApJ, 656, 847 Wang, Q. D., Lu, F., & Lang, C. C. 2002, ApJ, 581, 1148 Zhang, Z.-Y., Gao, Y., Henkel, C., et al. 2014, ApJ, 784, LL31 Zoglauer, A., Reynolds, S. P., An, H., et al. 2015, ApJ, 798, 98 c 0000 RAS, MNRAS 000, 1–??
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project