The origin of the chemical elements in cluster cores
aa r X i v : . [ a s t r o - ph . C O ] O c t Astronomische Nachrichten, 12 November 2018
The origin of the chemical elements in cluster cores
J. de Plaa ,⋆ SRON Netherlands Institute for Space Research, Sorbonnelaan 2, 3584 CA UtrechtReceived 2 July 2012, accepted XXXXPublished online XXXX
Key words
X-rays: galaxies: clusters, galaxies: clusters: general, galaxies: abundances, supernovae: generalMetals play a fundamental role in ICM cooling processes in cluster cores through the emission of spectral lines. But whenand how were these metals formed and distributed through the ICM? The X-ray band has the unique property of containingemission lines from all elements from carbon to zinc within the 0.1 −
10 keV band. Using XMM-Newton, the abundancesof about 11 elements are studied, which contain valuable information about their origin. Most elements were formedin type Ia and core-collapse supernovae, which have very different chemical yields. Massive stars and AGB stars alsocontribute by providing most of the carbon and nitrogen in the ICM. Because feedback processes suppress star formationin the cluster centre, the element abundances allow us to directly probe the star formation history of the majority of starsthat are thought to have formed between z = 2 − . The spatial distribution in the core and the evolution with redshiftalso provide information about how these elements are transported from the member galaxies to the ICM. I review thecurrent progress in chemical enrichment studies of the ICM and give an outlook to the future opportunities provided byXMM-Newton’s successors, like Astro-H. Copyright line will be provided by the publisher
After the ‘Big Bang’, the baryonic component of the Uni-verse mainly consisted of hydrogen and helium with tracesof lithium and beryllium. The first metals with a higheratomic weight were produced in the first generation of stars,also referred to as Population III stars, which started theepoch of re-ionization. The nature of this stellar popula-tion is very uncertain, but it likely consisted of intermediate-mass and high-mass stars (Vangioni et al. 2011). Based onWMAP measurements, these stars formed around z ∼ ,when the age of the Universe was about 500 Myr. Althoughthese stars were the first to enrich the surrounding gas, thetotal contribution to the current day chemical composition isthought to be small, only 10 − Z solar (Matteucci & Calura2005).The bulk of the enrichment probably occurred around z ∼ − through the supernova explosions followingmajor star bursts. A compilation of the universal star for-mation rate measurements as a function of redshift (Hop-kins & Beacom 2006) shows that the star formation ratepeaked around z ∼ − and declined slowly to low red-shift. However, these are averaged rates over both clusterand field galaxies. At a similar redshift, the Intra-ClusterMedium (ICM) in clusters of galaxies starts to form, which,together with feedback from Active Galactic Nuclei (AGN),quenched the star formation in these objects. Therefore, inthe cluster of galaxies case, the star formation rate dropsmuch faster than average (Gabor et al. 2010). In that respect,clusters of galaxies are a special environment, because their ⋆ Corresponding author: e-mail: [email protected] enrichment is dominated by the products from the main starbursts at z ∼ − .Ferrara et al. (2005) have estimated that a major fraction( > z = 3 contributesto the enrichment of gas in a hot phase, instead of ending innewly formed stars. The ICM in clusters therefore containsthe bulk of the metals produced in the cluster member galax-ies. In the period of the major star bursts, the main enrich-ment mechanisms that were transporting the metals fromthe galaxies to the surrounding medium were galactic windsdriven by the supernova explosions and AGN uplifting ofgalactic gas. At a lower level, also other enrichment mech-anisms play a role. Metals can be ejected from the galaxiesby galaxy-galaxy interactions, by sloshing motions of thehot ICM, and ram-pressure stripping of in-falling galaxies(see Schindler & Diaferio 2008 for a review). The bulk of the elements heavier than beryllium are pro-duced in the end phases of stellar evolution. The elements inthe mass range between oxygen and silicon are mainly pro-duced by core-collapse supernovae, because the elementswith masses higher than silicon that are produced in the coreregion of the massive star are compressed into the neutronstar or black hole that is formed during the supernova event.Only the outer layers containing the lighter elements areejected into the surrounding medium. The elements fromsilicon to nickel, however, are the main products of typeIa supernovae, which are exploding white dwarfs in binarysystems. During such a supernova event, elements up to
Copyright line will be provided by the publisher
J. de Plaa: The origin of the chemical elements in cluster cores
Fig. 1
Expected abundances measured in a 120 ks XMM-Newton observation of S´ersic 159-03 (bottom panel), whichis a typical bright local cluster. The statistical error barswere obtained from de Plaa et al. (2006). The estimatesfor the SNIa, SNcc, and AGB contributions are based ona sample of 22 clusters (de Plaa et al. 2007) and two ellip-tical galaxies (Grange et al. 2011). The top panels show thetypical range in SNIa and IMF models with respect to thestatistical error bars in the observation.nickel are produced by explosive fusion and the white dwarfis disintegrated, which releases all the products in the sur-rounding medium.A few light elements, however, have a different origin.Carbon is thought to be mainly produced by massive starsduring their lifetime and ejected through the stellar winds.Nitrogen is ejected by intermediate-mass stars in the Asymp-totic Giant Branch (AGB) phase. Since elements heavierthan nickel are not produced by fusion, and are thereforeless abundant, we do not discuss the origin of these metalsin this paper.Figure 1 shows the expected contributions of the mostabundant metals that can be detected using the XMM-New-ton observatory. The bars indicate the relative contributionof each supernova type to the abundance of the elements.The fractions depend on the supernova models used. Theeffect of uncertainties in the supernova models and in theInitial-Mass Function (IMF) are indicated in the top panel.
Although type Ia supernovae are used as standard candlesfor cosmology, their progenitor and explosion mechanismare still poorly known. The common misconception is thata type Ia supernova occurs when a white dwarf goes overthe Chandrasekhar mass. Instead, the explosive carbon fu-sion in the star is ignited just before it reaches this limit. Inrecent years, optical searches for supernovae have yieldedhundreds of observations of type Ia’s which revealed a sur-prising variety in their properties. It has proved to be verydifficult to link the observed supernovae to possible progen-itors (see e.g. Howell 2011 for a recent review). Currently, roughly three main progenitors are being con-sidered. The first is a ‘classical’ type Ia or also called ‘Sin-gle Degenerate’ (SD), where the companion of the whitedwarf is a main sequence or red-giant star that accretes ma-terial on the white dwarf. The second is a scenario wheretwo white dwarfs merge and the less massive white dwarfis accreted onto the more massive one, which is known asthe ‘Double Degenerate’ (DD) scenario. And the last is asub-Chandrasekhar channel, where a thick layer of heliumbuilds up on the white dwarf’s surface either by hydrogenburning or by a helium-rich donor. Because of the lowerdensity of the last systems, they are thought to produce moreintermediate-mass elements like silicon, sulfur, and calciumthan during the deflagration in the typical SD and DD cases.Observations have shown that relatively luminous typeIa supernovae tend to occur in spiral galaxies, while thesub luminous are mainly found in elliptical galaxies withold stellar populations (e.g. Howell 2001; Sullivan et al.2006). This appears to indicate that the brighter ’prompt’supernovae explode early ( <
400 Myr) after the star burst,while the sub-luminous occur with a delay of a few Gyr.This can be explained in the DD scenario by the fact thatit takes a longer time to evolve a star into a low-mass whitedwarf compared to a high-mass white dwarf. The delay timeexplanation is also supported by high-redshift studies thatshow that at z = 1 type Ia supernovae are 12% more lu-minous and have less intermediate mass elements in theirspectra than local SNIa’s (e.g. Howell et al. 2007; Sullivanet al. 2009). However, also differences in initial metallicitiesor other unknown parameters could play a role.One of the few successes to combine theory and obser-vations in this field is the explanation of the observed Delay-Time Distribution (DTD) using stellar population synthesisstudies. Supernova rate measurements in elliptical galaxies(and clusters of galaxies) show that the rate declines with ∼ t − after the initial star burst. Model calculations showthe same level of decline for DD mergers, which is an indi-cation that this channel is dominating the SNIa rate at largerdelay times (see Fig 2; Ruiter et al. 2009; Mennekens et al.2010).Due to the uncertainty in the nature of the SNIa progeni-tor and the poorly understood physics of the explosion (dur-ing the explosion, the burning front probably switches frombeing subsonic to supersonic), there is a large range of su-pernova Ia models predicting the metal yields. Well knownexamples are Iwamoto et al. (1999) and models by Bravoet al. (2012). The predicted yields for various elements canvary substantially for every model, which means that ac-curate abundance measurements can help to constrain theSNIa models. Although there is also a range of different models for core-collapse supernova yields in the literature (e.g. Woosley &Weaver, 1995; Tsujimoto et al., 1995; Chieffi & Limongi,
Copyright line will be provided by the publisher
N header will be provided by the publisher 3
Fig. 2
Supernova type Ia delay time distributions for threedifferent progenitor channels: white-dwarf mergers (DDS),Single-Degenerate Systems (SDS), and AM CVn systemsas described in Ruiter et al. (2009). AM CVn systems areultra-compact systems with a WD accretor and a He-richdonor. The top and bottom panel represent the result fromtwo choices of initial parameters, but both are calculatedfor stellar populations in elliptical galaxies. The dotted lineshows the observed decline of the type Ia supernova rate of ∼ t − , where t is the time passed since the initial star burst.The decline is consistent with the DDS progenitor model.2004), the main parameters that determine the abundancesmeasured are initial metallicity and the choice of the IMF.In order to determine the total contribution of core-collapsesupernovae to the cluster enrichment, the model yields foreach mass bin need to be weighted with the IMF in the fol-lowing way (Tsujimoto et al, 1995): M i = R
50 M ⊙
10 M ⊙ M i ( m ) m − (1+ x ) d m R
50 M ⊙
10 M ⊙ m − (1+ x ) d m , (1)where M i ( m ) is the i th element mass produced in a main-sequence star of mass m . A standard Salpeter IMF corre-sponds to x = 1 . , but this equation can, of course, bealtered to represent other IMFs. Since intermediate-mass stars in their AGB phase are a ma-jor source of nitrogen, it is important to include them in the
Fig. 3
EPIC spectrum of the cluster 2A 0335+096 withan exposure time of 130 ks. Adapted from Werner et al.(2006b).model as well. Since the yields depend on the initial massof the main sequence star, we need to weigh the yields alsowith the IMF like in Equation 1. AGB star yields for differ-ent masses and initial metallicities are calculated by, for ex-ample, Karakas (2010). Together with the information fromthe core-collapse models, the AGB stars provide an addi-tional constraint to the initial metallicity and IMF of thestellar population.
The soft X-ray band between 0.1 and 10 keV is very suit-able to measure abundances in hot plasmas, because it con-tains spectral lines of all elements from carbon to nickel.The hot ICM in clusters is particularly interesting, becausethe plasma is in collisional-ionization equilibrium, whichsimplifies the determination of the abundances. Moreover,the cluster ICM contains the integrated yield of billions ofsupernovae, providing a general picture of supernova yields,contrary to studies of a few individual supernovae throughtheir remnants in our galaxy. The sensitivity of XMM-New-ton, and also Suzaku, allows us to measure abundances withenough accuracy to constrain supernova models. An exam-ple of a deep EPIC observation of the cluster 2A 0335+096in Figure 3 shows the position and strength of the lines of themost abundant elements. The availability of these excellentdata sets has triggered a number of successful abundancestudies (see also the review by Werner et al. 2008).
The first attempt to link measured abundances in the hotICM in local clusters ( z < . ) was performed using theASCA satellite (Mushotzky et al., 1996). Its instruments al-lowed the accurate detection of O, Ne, Mg, Si, S, Ar, Ca, Fe,and Ni for the first time. From this and later ASCA studies Copyright line will be provided by the publisher
J. de Plaa: The origin of the chemical elements in cluster cores
Fig. 4
Abundance ratios fitted with supernova yields fromthe WDD2 SNIa model (Iwamoto et al. 1999) and a SNccmodel with an initial metallicity Z = 0 . and a SalpeterIMF. The calcium abundance appears to be underestimated.It can not explain the Ar/Ca ratio measured in this XMM-Newton sample of 22 clusters. (Adapted from de Plaa et al.2007).(e.g. Fukazawa et al. 1998; Finoguenov et al. 2000) a gen-eral picture emerged of an ICM which was enriched earlyin the cluster evolution by core-collapse supernova and laterby delayed type Ia supernovae.When more deep observations of clusters with XMM-Newton became available, de Plaa et al. (2007) performedan abundance study in a sample of 22 clusters with EPIC.With the obtained accuracy, it was possible to estimate therelative contributions of type Ia and core-collapse super-nova to the enrichment of the cluster by linearly combiningthe supernova yields for each type. For the most commonlyused type Ia models (Iwamoto et al. 1999), this results in aSNIa contribution of about 30%, which is remarkably simi-lar to the SNIa/SNcc ratio of ∼ . − . as observed in theoptical band (e.g. Horiuchi et al. 2011) given the fact thatde Plaa et al. (2007) ignored galactic evolution effects. Sur-prisingly, the calcium abundance did not fit the Iwamoto etal. (1999) models well (see Fig 4), while a model by Bravoet al. (1996), which also fitted to the Tycho supernova rem-nant (Badenes et al. 2006), resulted in a much better fit ofthe Ar/Ca ratio (de Plaa et al. 2007). This showed that X-rayspectra enable us to constrain type Ia supernova models.Several groups have performed similar studies with XMM-Newton and Suzaku data since then, but usually on fewerclusters and with varying numbers of elements. In a smallsample, Sato et al. (2007) performed the same fit as de Plaaet al. (2007) using abundances measured with Suzaku XIS.Unfortunately, argon and calcium were not included in thefit, but their main conclusion was consistent with de Plaaet al. (2007) considering the uncertainty introduced by thelimited number of measured elements. Recently, Bulbul etal. (2012) used a modified APEC model that is able to fit the supernova type Ia to core-collapse ratio directly to thespectra. They also report a type Ia contribution of 30–40%,which is consistent with previous work and optical data. Itshould be noted that this ratio depends highly on the as-sumed supernova models (De Grandi & Molendi, 2009), butit is reassuring that different groups find similar numbers in-dependently. In recent years, deep XMM-Newton observations of giantelliptical galaxies also allowed the detection of carbon andnitrogen. These two elements are not produced in large quan-tities in core-collapse supernovae, but in a variety of sources.The main origins of these elements are still subject of de-bate (see e.g. Romano et al. 2010), but likely candidatesare metal-poor massive stars and intermediate-mass stars intheir AGB phase.Since carbon and nitrogen have a relatively low atomicmass, their K-shell transitions are located in the soft X-rayband below 0.5 keV. It also causes the lines to be stronger atlower temperatures < keV (at temperatures of a few keV,a larger fraction of the carbon and nitrogen is fully ionisedand not producing line emission). To detect the lines, it isbest to observe cool elliptical galaxies with the ReflectionGrating Spectrometer (RGS) aboard XMM-Newton. Thisinstrument has sufficient effective area at low energies anda high spectral resolution to resolve the lines of these el-ements. Unfortunately, noise in the RGS CCD in the bandwhere the carbon line is located prevents an accurate carbonmeasurement for weak sources.Werner et al. (2006a) and Grange et al. (2011) anal-ysed RGS data of the giant elliptical galaxies M87 (Werner),NGC 5044, and NGC 5813 (Grange) and detected nitrogen.They found that the nitrogen abundances in these objects areindeed much higher than expected based on supernova mod-els. From the high N/O ratio, it can be deduced that nitro-gen should still be produced in low- and intermediate-massstars contrary to supernovae, because the star-formation ratein elliptical galaxies has declined very rapidly after the starbursts around z ∼ − . Studying the chemical enrichment of clusters above z = 0 . becomes increasingly difficult, because of the lower fluxof the sources. The Fe-K feature is usually the only fea-ture strong enough to resolve at higher redshifts. Therefore,studies at these large distances focus on the iron abundanceevolution as a function of redshift. In a sample of 56 clus-ters with redshifts in the 0.3 < z < z = 0 . to 0.2 solar at z = 1 . . It was, however, also clear that the scatter onthe measured data points was relatively large. In a differentsample of 39 clusters observed with XMM-Newton, Baldi Copyright line will be provided by the publisher
N header will be provided by the publisher 5
Fig. 5
Measured abundances versus redshift for a sampleof intermediate redshift clusters. The top panel shows theabundances measured up to 0.6R and the bottom panelthe results when the core of the cluster is ignored. Adaptedfrom Baldi et al. (2012).et al. (2012) recently reported that they could not confirmthe trend found by Balestra et al. (2007), because of thehigh level of scatter in the data points (see Figure 5). Re-moving the cool cores in the spectral analysis just had alimited effect on the amount of scatter and did not allowthem to draw firm conclusions. Since these measurementsstill only sample the tail of the SNIa delay-time distribu-tion that originates at z = 2 − , a strong trend may notbe expected. Clearly, instruments with a much higher effec-tive area are needed to reduce the statistical errors on thesemeasurements. X-ray spectroscopy of clusters of galaxies has proved to bea useful tool to study the enrichment history of the hot ICM,but the field is still in its early stages of development. Usingcurrent instruments, it is possible to measure the abundanceof about 10 elements with an accuracy of 20–30% if oneincludes systematic errors (De Grandi & Molendi, 2009).This appears to be a substantial systematic error, however,the uncertainty in the supernova yields, given by the spreadin supernova models, is sometimes more than a factor of twofor certain elements. This means that even with these sys-tematic uncertainties, the measurements can put constraintson supernova models and the IMF.There are two main sources of systematic errors in abun-dance measurements: the atomic data in spectroscopic codesand instrument calibration issues. A third one can be intro- . . C oun t s / s / k e V Energy (keV)Abell 1795 (300 ks)XMM−Newton Astro−H SXS
C N O Fe−L
Ne Na
Mg Al Si P SCl Ar K Ca Ti Cr Mn Fe−K Co Ni CuZn
EPIC MOS1
Fig. 6
Simulated Astro-H micro-calorimeter spectrum ofAbell 1795 compared to an XMM-Newton EPIC MOSspectrum. The spectra were calculated using an updated ver-sion of the atomic database in SPEX including trace ele-ments, like Cl, K, Ti, Cr, and Mn.duced when the multi-temperature structure in the gas isnot taken properly into account in the spectral fit (Buote& Fabian 1998). Both the systematics in the atomic dataand the calibration depend very much on the specific tran-sition and wavelength of the line. The silicon abundance,for example, is mainly determined by measuring the Si XIVlines at about 2.0 keV, which is very close to the Au-edgeof the mirror and Si-edge of the detector. This is a verychallenging wavelength band to calibrate accurately. In thiscase, the Si XIV lines are well characterized in the atomicdatabase, while there are also other lines from, for example,Fe XVII that have a much larger uncertainty in their oscil-lator strengths (de Plaa et al., 2012). It is very important tounderstand the systematics for each measured element indi-vidually to assess the quality of the supernova fits.In general, most of the measured abundances have a sys-tematic uncertainty that is lower than or not much higherthan 20%, which is still much better than the scatter in themodels. To beat down these systematic errors, investmentshave to be made in the development of spectral codes andtheir underlying atomic data. In addition, a high level of cal-ibration accuracy is to be pursued to improve the abundanceaccuracy both for current and future missions.
The next leap forward will be made using micro-calorimetertechnology, which will allow spatially resolved high-reso-lution X-ray spectroscopy of extended sources. This typeof sensor, which will be flown on Astro-H (Takahashi etal. 2010), has a typical resolution of a few eV across thesoft X-ray band. With this technology, it will be possibleto resolve lines from less abundant elements, like sodium,chromium, and manganese (see Fig 6). Knowledge about
Copyright line will be provided by the publisher
J. de Plaa: The origin of the chemical elements in cluster cores the abundances of these trace elements puts additional con-straints on supernova models. The Mn/Cr ratio, for example,appears to be very sensitive to the metallicity of the type Iaprogenitor (Badenes et al. 2008) and the Na abundance is,like nitrogen, also correlated with the contribution of inter-mediate mass AGB stars and sensitive to the IMF.To measure trace elements like chromium and manga-nese in high resolution spectra, the spectral fitting codesthat are used to fit the spectra need updates to their atomicdata. Until recently, only the lines from elements with a highabundance (typically with even atomic numbers) where tak-en into account in the codes. The commonly used databaseslike ATOMDB (Smith et al. 2001) and the SPEX linedatabase (Kaastra et al. 1996) are being updated and are(partly) available in the spectral fitting codes. The system-atic errors in the atomic data, which can reach the 20% levelin some important lines, remain an issue. Further invest-ments in laboratory measurements of lines and theoreticalline calculations are needed to improve the accuracy.Until now, the measured abundances have been com-pared to individual supernova models. It would, however,be much more realistic to estimate the expected abundancesfrom advanced binary population synthesis codes. Using thesecodes, it is possible to give the proper weight to the differenttype Ia progenitor scenarios and their respective yields. Thehigh calcium abundance measured by de Plaa et al. (2007),for example, might be explained by sub-Chandrasekhar typeIa’s, because explosive helium fusion produces more inter-mediate mass elements, like calcium. Predicted abundancesusing these detailed binary population synthesis can be di-rectly compared to the measured abundances in clusters.The high amount of measurable elements will substantiallyimprove the quality of the test and therefore also our knowl-edge of chemical enrichment in clusters. Acknowledgements.
The author likes to thank Jelle Kaastra forcarefully reading the manuscript and useful discussions. SRONNetherlands Institute for Space Research is supported financiallyby NWO, the Netherlands Organisation for Scientific Research.
References
Badenes, C., Borkowski, K. J., Hughes, J. P., Hwang, U., Bravo,E.: 2006, ApJ 645, 1373Badenes, C., Bravo, E., Hughes, J. P.: 2008, ApJ 680, L33Baldi, A., Ettori, S., Molendi, S., Balestra, I., Gastaldello, F.,Tozzi, P.: 2012, A&A 537, A142Balestra, I., Tozzi, P., Ettori, S., Rosati, P., Borgani, S., Mainieri,V., Norman, C., Viola, M.: 2007, A&A 462, 429Bravo, E., Mart´ınez-Pinedo, G.: 2012, PhRvC 85(5), 055805Bravo, E., Tornambe, A., Dominguez, I., Isern, J.: 1996, A&A 306,811Bulbul, E., Smith, R. K., Loewenstein, M.: 2012, ApJ 753, 54Buote, D. A., Fabian, A. C.: 1998, MNRAS 296, 977Chieffi, A., Limongi, M.: 2004, ApJ 608, 405de Grandi, S., Molendi, S.: 2009, A&A 508, 565 de Plaa, J., Werner, N., Bleeker, J. A. M., Vink, J., Kaastra, J. S.,M´endez, M.: 2007, A&A 465, 345de Plaa, J., Werner, N., Bykov, A. M., Kaastra, J. S., M´endez, M.,Vink, J., Bleeker, J. A. M., Bonamente, M., Peterson, J. R.:2006, A&A 452, 397de Plaa, J., Zhuravleva, I., Werner, N., Kaastra, J. S., Churazov, E.,Smith, R. K., Raassen, A. J. J., Grange, Y. G.: 2012, A&A 539,A34Ferrara, A., Scannapieco, E., Bergeron, J.: 2005, ApJ 634, L37Finoguenov, A., David, L. P., Ponman, T. J.: 2000, ApJ 544, 188Fukazawa, Y., Makishima, K., Tamura, T., Ezawa, H., Xu, H.,Ikebe, Y., Kikuchi, K., Ohashi, T.: 1998, PASJ 50, 187Gabor, J. M., Dav´e, R., Finlator, K., Oppenheimer, B. D.: 2010,MNRAS 407, 749Grange, Y. G., de Plaa, J., Kaastra, J. S., Werner, N., Verbunt, F.,Paerels, F., de Vries, C. P.: 2011, A&A 531, A15Hopkins, A. M., Beacom, J. F.: 2006, ApJ 651, 142Horiuchi, S., Beacom, J. F., Kochanek, C. S., Prieto, J. L., Stanek,K. Z., Thompson, T. A.: 2011, ApJ 738, 154Howell, D. A.: 2001, ApJ 554, L193Howell, D. A.: 2011, Nature Communications 2, 350Howell, D. A., Sullivan, M., Conley, A., Carlberg, R.: 2007,ApJ 667, L37Iwamoto, K., Brachwitz, F., Nomoto, K., Kishimoto, N., Umeda,H., Hix, W. R., Thielemann, F. K.: 1999, ApJS 125, 439Kaastra, J. S., Mewe, R., Nieuwenhuijzen, H.: 1996, SPEX: a newcode for spectral analysis of X & UV spectra. In Yamashita,K., Watanabe, T. (Eds.), UV and X-ray Spectroscopy of As-trophysical and Laboratory Plasmas (pp. 411–414)Karakas, A. I.: 2010, MNRAS 403, 1413Matteucci, F., Calura, F.: 2005, MNRAS 360, 447Mennekens, N., Vanbeveren, D., De Greve, J. P., De Donder, E.:2010, A&A 515, A89Mushotzky, R., Loewenstein, M., Arnaud, K. A., Tamura, T.,Fukazawa, Y., Matsushita, K., Kikuchi, K., Hatsukade, I.:1996, ApJ 466, 686Romano, D., Karakas, A. I., Tosi, M., Matteucci, F.: 2010, A&A522, A32Ruiter, A. J., Belczynski, K., Fryer, C.: 2009, ApJ 699, 2026Sato, K., Tokoi, K., Matsushita, K., Ishisaki, Y., Yamasaki, N. Y.,Ishida, M., Ohashi, T.: 2007, ApJ 667, L41Schindler, S., Diaferio, A.: 2008, SSRv 134, 363Smith, R. K., Brickhouse, N. S., Liedahl, D. A., Raymond, J. C.:2001, ApJ 556, L91–L95Sullivan, M., Ellis, R. S., Howell, D. A., Riess, A., Nugent, P. E.,Gal-Yam, A.: 2009, ApJ 693, L76Sullivan, M., Le Borgne, D., Pritchet, C. J., Hodsman, A., Neill,J. D., Howell, D. A., Carlberg, R. G., et al.: 2006, ApJ 648,868Takahashi, T., Mitsuda, K., Kelley, R., et al.: 2010, The ASTRO-HMission. In Proc. SPIE, Volume 7732 (p. 77320Z)Tsujimoto, T., Nomoto, K., Yoshii, Y., Hashimoto, M., Yanagida,S., Thielemann, F. K.: 1995, MNRAS 277, 945Vangioni, E., Silk, J., Olive, K. A., Fields, B. D.: 2011, MN-RAS 413, 2987Werner, N., B¨ohringer, H., Kaastra, J. S., de Plaa, J., Simionescu,A., Vink, J.: 2006a, A&A 459, 353Werner, N., de Plaa, J., Kaastra, J. S., Vink, J., Bleeker, J. A. M.,Tamura, T., Peterson, J. R., Verbunt, F.: 2006b, A&A 449, 475Werner, N., Durret, F., Ohashi, T., Schindler, S., Wiersma, R. P. C.:2008, SSRv 134, 337Woosley, S. E., Weaver, T. A.: 1995, ApJS 101, 181de Plaa, J., Werner, N., Bleeker, J. A. M., Vink, J., Kaastra, J. S.,M´endez, M.: 2007, A&A 465, 345de Plaa, J., Werner, N., Bykov, A. M., Kaastra, J. S., M´endez, M.,Vink, J., Bleeker, J. A. M., Bonamente, M., Peterson, J. R.:2006, A&A 452, 397de Plaa, J., Zhuravleva, I., Werner, N., Kaastra, J. S., Churazov, E.,Smith, R. K., Raassen, A. J. J., Grange, Y. G.: 2012, A&A 539,A34Ferrara, A., Scannapieco, E., Bergeron, J.: 2005, ApJ 634, L37Finoguenov, A., David, L. P., Ponman, T. J.: 2000, ApJ 544, 188Fukazawa, Y., Makishima, K., Tamura, T., Ezawa, H., Xu, H.,Ikebe, Y., Kikuchi, K., Ohashi, T.: 1998, PASJ 50, 187Gabor, J. M., Dav´e, R., Finlator, K., Oppenheimer, B. D.: 2010,MNRAS 407, 749Grange, Y. G., de Plaa, J., Kaastra, J. S., Werner, N., Verbunt, F.,Paerels, F., de Vries, C. P.: 2011, A&A 531, A15Hopkins, A. M., Beacom, J. F.: 2006, ApJ 651, 142Horiuchi, S., Beacom, J. F., Kochanek, C. S., Prieto, J. L., Stanek,K. Z., Thompson, T. A.: 2011, ApJ 738, 154Howell, D. A.: 2001, ApJ 554, L193Howell, D. A.: 2011, Nature Communications 2, 350Howell, D. A., Sullivan, M., Conley, A., Carlberg, R.: 2007,ApJ 667, L37Iwamoto, K., Brachwitz, F., Nomoto, K., Kishimoto, N., Umeda,H., Hix, W. R., Thielemann, F. K.: 1999, ApJS 125, 439Kaastra, J. S., Mewe, R., Nieuwenhuijzen, H.: 1996, SPEX: a newcode for spectral analysis of X & UV spectra. In Yamashita,K., Watanabe, T. (Eds.), UV and X-ray Spectroscopy of As-trophysical and Laboratory Plasmas (pp. 411–414)Karakas, A. I.: 2010, MNRAS 403, 1413Matteucci, F., Calura, F.: 2005, MNRAS 360, 447Mennekens, N., Vanbeveren, D., De Greve, J. P., De Donder, E.:2010, A&A 515, A89Mushotzky, R., Loewenstein, M., Arnaud, K. A., Tamura, T.,Fukazawa, Y., Matsushita, K., Kikuchi, K., Hatsukade, I.:1996, ApJ 466, 686Romano, D., Karakas, A. I., Tosi, M., Matteucci, F.: 2010, A&A522, A32Ruiter, A. J., Belczynski, K., Fryer, C.: 2009, ApJ 699, 2026Sato, K., Tokoi, K., Matsushita, K., Ishisaki, Y., Yamasaki, N. Y.,Ishida, M., Ohashi, T.: 2007, ApJ 667, L41Schindler, S., Diaferio, A.: 2008, SSRv 134, 363Smith, R. K., Brickhouse, N. S., Liedahl, D. A., Raymond, J. C.:2001, ApJ 556, L91–L95Sullivan, M., Ellis, R. S., Howell, D. A., Riess, A., Nugent, P. E.,Gal-Yam, A.: 2009, ApJ 693, L76Sullivan, M., Le Borgne, D., Pritchet, C. J., Hodsman, A., Neill,J. D., Howell, D. A., Carlberg, R. G., et al.: 2006, ApJ 648,868Takahashi, T., Mitsuda, K., Kelley, R., et al.: 2010, The ASTRO-HMission. In Proc. SPIE, Volume 7732 (p. 77320Z)Tsujimoto, T., Nomoto, K., Yoshii, Y., Hashimoto, M., Yanagida,S., Thielemann, F. K.: 1995, MNRAS 277, 945Vangioni, E., Silk, J., Olive, K. A., Fields, B. D.: 2011, MN-RAS 413, 2987Werner, N., B¨ohringer, H., Kaastra, J. S., de Plaa, J., Simionescu,A., Vink, J.: 2006a, A&A 459, 353Werner, N., de Plaa, J., Kaastra, J. S., Vink, J., Bleeker, J. A. M.,Tamura, T., Peterson, J. R., Verbunt, F.: 2006b, A&A 449, 475Werner, N., Durret, F., Ohashi, T., Schindler, S., Wiersma, R. P. C.:2008, SSRv 134, 337Woosley, S. E., Weaver, T. A.: 1995, ApJS 101, 181