A solar-type star polluted by calcium-rich supernova ejecta inside the supernova remnant RCW 86

TL;DR

This paper reports the discovery of a solar-type star polluted by calcium-rich supernova ejecta inside the supernova remnant RCW 86, providing evidence that calcium-rich supernovae can originate from core-collapse explosions.

Contribution

It presents the first observational evidence linking calcium-rich supernovae to core-collapse origins through pollution in a binary system within a supernova remnant.

Findings

Solar-type star strongly polluted with calcium and other elements.

Supernova progenitor was a moving star that lost mass before explosion.

Calcium-rich supernovae can arise from core-collapse explosions.

Abstract

When a massive star in a binary system explodes as a supernova its companion star may be polluted with heavy elements from the supernova ejecta. Such a pollution had been detected in a handful of post-supernova binaries (Gonzalez Hernandez et al. 2011), but none of them is associated with a supernova remnant. We report the discovery of a solar-type star in a close, eccentric binary system with a neutron star within the young Galactic supernova remnant RCW 86. Our discovery implies that the supernova progenitor was a moving star, which exploded near the edge of its wind bubble and lost most of its initial mass due to common-envelope evolution shortly before core collapse. We find that the solar-type star is strongly polluted with calcium and other elements, which places the explosion within the class of calcium-rich supernovae -- faint and fast transients (Filippenko et al. 2003; Kasliwal et al. 2012}, whose origin is strongly debated (Kawabata et al. 2010; Waldman et al. 2011), and provides the first observational evidence that supernovae of this type can arise from core-collapse explosions (Kawabata et al. 2010; Moriya et al. 2010).

Tables

Table 1. Table 1: Radial velocity changes with time in the spectrum of [GV2003] N

Date	Radial	Heliocentric
	velocity∗	correction
	($\mathrm{km}{s}^{-1}$)	($\mathrm{km}{s}^{-1}$)
2015 April 14	$-71$$\pm$2	13.0
2015 April 21	$-74$$\pm$4	10.9
2015 May 13	$-44$$\pm$3	3.3
2015 May 16	$-62$$\pm$3	2.4

Table 2. Table 2: Extended Data Table 1 Surface element abundances of [GV2003] N

Element	$X/H$	$M_{\mathrm{acc}}/\left({10}^{-5}{M}_{\odot }\right)$
Li	0.29∗	–
Si	0.24$\pm$0.10(0.13)†	8
Ca	0.81$\pm$0.05(0.24)	7
Ti	0.50$\pm$0.18(0.46)	0.09
V	0.67$\pm$0.01(0.37)	0.028
Cr	0.47$\pm$0.10(0.23)	0.6
Mn	0.62$\pm$0.18(0.27)	0.8
Fe	0.22$\pm$0.10(0.19)	15
Co	0.38$\pm$0.04(0.23)	0.09
Ni	0.45$\pm$0.33(0.35)	1.8
Ba	$-$0.04$\pm$0.10(0.43)	0

Full text

{affiliations}

Sternberg Astronomical Institute, Lomonosov Moscow State University, Universitetskij Pr. 13, Moscow 119992, Russia

Space Research Institute, Russian Academy of Sciences, Profsoyuznaya 84/32, 117997 Moscow, Russia

Isaac Newton Institute of Chile, Moscow Branch, Universitetskij Pr. 13, Moscow 119992, Russia

Argelander-Institut für Astronomie, Auf dem Hügel 71, 53121 Bonn, Germany

Space Research Institute, Austrian Academy of Sciences, Schmiedlstrasse 6, 8042 Graz, Austria

CSIRO Astronomy and Space Science, Australia Telescope National Facility, PO Box 76, Epping, NSW 1710, Australia

Department of Astronomy, University of Michigan, 1085 S. University Avenue, Ann Arbor, MI 48109, USA

Max-Planck Institut für Astronomie, Königstuhl 17, 69117 Heidelberg, Germany

Max-Planck-Institut für extraterrestrische Physik, Giessenbachstr. 1, 85748 Garching, Germany

Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, 53121 Bonn, Germany

Recently, it was recognized that a large fraction of massive stars reside in binary systems[7, 8] and that the evolution of the majority of massive stars is strongly affected by binary interaction, through mass transfer, common envelope evolution or merger[7]. This suggests that most Type Ib/c supernovae (SNe), which do not show hydrogen lines in their spectra, stem from progenitors which are stripped of their hydrogen envelopes by a companion star [10, 9]. The low ejecta masses typical of such SNe[11] imply that a significant fraction of the post-SN binaries remain bound. Only very few of such binaries are known[12, 13] to be associated with supernova remnants (SNRs) because of the short lifetime (\sim$$10^{5} yr) of the SNRs. In this Letter, we report the discovery of a post-SN binary system within the young (few thousand years[14]) SNR RCW 86.

The pyriform appearance of RCW 86 (Fig. 1; see also fig. 6 in ref.[16]) can be explained as the result of a SN explosion near the edge of a bubble blown by the wind of a moving massive star[17, 18] (Supplementary Information section 1). This interpretation implies that the SN exploded near the centre of the hemispherical optical nebula in the south-west of RCW 86 (see Fig. 1) and that the stellar remnant should still be there. Motivated by these arguments, we looked for a possible compact X-ray source using archival Chandra data and discovered[17] two sources in the expected position of the SN progenitor (Fig. 1). One of them, [GV2003] S, has a clear optical counterpart with $V$=14.4 mag and its X-ray spectrum implies that this source is a foreground late-type active star. For the second source, [GV2003] N, we did not find any optical counterpart in the Digital Sky Survey II to a limiting red band magnitude of $\approx$21, while its X-ray spectrum suggests that this source could be a young pulsar[17]. Our deep follow-up observation with the Parkes radio telescope in 2002, however, failed to detect any radio emission from [GV2003] N, giving an upper limit on the flux of 35 $\mu$Jy at 1420 MHz (see Methods). This non-detection may be a consequence of beaming or it could indicate that [GV2003] N may not be an active radio pulsar.

If [GV2003] N was a NS its emission in the visual was expected to be fainter than V$$\approx28 mag. We therefore obtained a $V$-band image of the field around this source with the FORS2 instrument on the ESO Very Large Telescope (VLT) in 2010. The FORS2 image, however, revealed a stellar-like object with $V$=20.69$\pm$0.02 mag just at the position of [GV2003] N (Fig. 1; Methods). To further constrain the nature of [GV2003] N, we obtained its $g^{\prime}r^{\prime}i^{\prime}z^{\prime}JHK_{\rm s}$ photometry with the 7-channel optical/near-infrared imager GROND in 2013 (Methods). With that, we fitted the spectral energy distribution (SED) of [GV2003] N and derived a temperature of $\approx$5200 K and a colour excess of E(B-V)$$\approx0.9 mag (Methods; Extended Data Fig. 1). These results exclude the possibility that [GV2003] N is an AGN (as it was proposed in ref.[19]) and strongly suggest that the optical emission originates from a G-type star at a distance comparable to that of RCW 86 of 2.3$\pm$0.2 kpc[20]. Since the X-ray luminosity of [GV2003] N of \sim$$10^{32}\,{\rm erg}\,{\rm s}^{-1} (ref.[17, 19]) is far too high for a G star[21], we arrived at the possibility that we are dealing with a G star orbiting the NS.

Consequently, we searched for radial velocity (RV) variability and traces of the SN ejecta in the spectrum of the optical counterpart of [GV2003] N. We obtained four spectra with the VLT/FORS2 in 2015. We found clear RV variations (Table 1), indicative of an eccentric binary with a period of about a month or less (Methods). The spectrum of [GV2003] N (Fig. 2) clearly resembles that of a solar-type star, confirming the results from the SED analysis.

From the analysis of the Fe i lines (Methods), we derived a stellar effective temperature of T_{\rm eff}=5100$$\pm200 K. Using $V$=20.69 mag, $E(B-V)$=0.9 mag and assuming the ratio of total to selective extinction of $R_{V}$=3.1, we estimated a visual extinction and absolute magnitude of $A_{V}$=2.79 mag and $M_{V}$=6.09 mag, respectively. With a bolometric correction of $-$0.29 mag[22], we obtained a bolometric luminosity of $\log(L/\rm\,L_{\odot})=-$0.42$\pm$0.08, taking into account the uncertainty on the distance. Using the Pisa stellar models[23], we estimated the mass of the optical star as $\approx 0.9\,\rm\,M_{\odot}$, which for the given $T_{\rm eff}$ and $L$ corresponds to a surface gravity of 4.6$\pm$0.1. The fact that the binary system remained bound after the SN explosion, and assuming the standard mass of a NS of $1.4\,\rm\,M_{\odot}$, implies a mass of the SN ejecta of less than $2.3\,\rm\,M_{\odot}$ — provided that the SN explosion was symmetric[24] — and a mass of the exploding star below $3.7\,\rm\,M_{\odot}$. This in turn implies that the SN explosion was of Type Ib and that the initial mass of the primary star was smaller than $13\,\rm\,M_{\odot}$[25].

From the FORS2 spectra we derived the abundances of Si, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, and Ba (Methods). Fig. 3 shows that many elements are enhanced by a factor of about 3 with respect to the solar abundances[26], with the silicon and iron being less than doubled. Calcium is particularly overabundant, by a factor of $\approx$6, which, to our knowledge, makes [GV2003] N the most Ca-rich star known to date.

The SN ejecta captured by the G star are mixed with the material of its convective envelope. Adopting an envelope mass of $0.2\,\rm\,M_{\odot}$[27], we computed the mass accreted from the SN ejecta for each element with measured abundances (Extended Data Table 1), which sums up to $3.3\times 10^{-4}\,\rm\,M_{\odot}$. Since the accreted mass of iron of \approx$$1.5\times 10^{-4}\,\rm\,M_{\odot} was ejected in the form of 56Ni, an accreted fraction of the SN ejecta of about 1% would lead to a total mass of radioactive nickel produced by the SN of $0.015\,\rm\,M_{\odot}$. These numbers appear consistent, since the explosion energy of SNe of \sim$$10\,\rm\,M_{\odot} stars is thought to be low, with the consequence of a relatively low nickel mass and moderate ejecta velocities[28], allowing for a rather high accreted fraction of the SN ejecta (Supplementary Information section 4).

The large amount of Ca in [GV2003] N, together with the rather modest enhancement of Si suggest that the SN, which produced RCW 86, was not a typical core-collapse event (Supplementary Information section 2). Instead, our data suggest that it belonged to the class of Ca-rich SNe — the fast and faint transients, which show strong Ca lines in their spectra and are believed to be intrinsically calcium-rich[29] (Supplementary Information section 3). Whereas observations indicate that many of the Ca-rich SNe are produced by long-lived low-mass stars, our findings imply that the lowest mass stars capable of producing NSs can produce similar explosion, if the progenitor is stripped of its envelope by a companion star. Indeed, from theoretical considerations, in both cases the explosion is produced in a helium layer with an inner radius comparable to that of a white dwarf, and an explosion kinetic energy of less than $10^{51}$ erg (Supplementary Information section 3).

The short orbital period of [GV2003] N implies that this binary system will evolve into a low-mass X-ray binary (LMXB) within its nuclear time scale (\sim$$10^{10} yr), providing the first definite example of a pre-LMXB located within a SNR. Before that, thermohaline mixing is expected to dilute the accreted SN material in the bulk of the mass of [GV2003] N on a time scale of \sim$$10^{7} yr. Thereby, the enhanced abundances of the measured elements will be reduced by factor of several, making them comparable to those found in the majority of the few main sequence components of LMXBs for which pollution has been detected previously[1].