A massive white-dwarf merger product before final collapse

TL;DR

This paper reports the discovery of a hot, rapidly rotating, highly magnetized white dwarf merger product with a nebula, providing evidence that some super-Chandrasekhar white dwarf mergers can avoid supernova explosions and evolve into neutron stars.

Contribution

It presents the first observational evidence of a massive white dwarf merger remnant with extreme properties, supporting models of super-Chandrasekhar mergers avoiding Type Ia supernovae.

Findings

Discovery of a hydrogen- and helium-free nebula with a hot central star.

The star has a surface temperature of about 200,000 K and outflow velocity of 16,000 km/s.

The star's properties match models of post-merger evolution of super-Chandrasekhar mass white dwarfs.

Abstract

Gravitational wave emission can lead to the coalescence of close pairs of compact objects orbiting each other. For the case of neutron stars such mergers may yield masses above the Tolman-Oppenheimer-Volkoff limit, leading to the formation of black holes. For the case of white dwarfs the merger product may exceed the Chandrasekhar limit, leading either to a thermonuclear explosion as Type Ia supernova, or to a collapse forming a neutron star. If a Type Ia supernova explosion is avoided, the merger of two massive white dwarfs is expected to form a hydrogen- and helium-free circumstellar nebula with a hot and luminous, rapidly rotating and highly magnetized central star for several 10,000 yr before its final collapse. Here we report the discovery of a hot star with an emission line dominated spectrum in the centre of a circular mid-infrared nebula. Both the star and the nebula appear to be free of hydrogen and helium. Our tailored stellar atmosphere and wind models indicate a stellar surface temperature of about 200,000 K, and a record outflow velocity of 16,000 km/s. This extreme velocity, together with the derived mass outflow rate, imply rapid stellar rotation and a strong magnetic field aiding the wind acceleration. The Gaia distance of the star leads to a luminosity of 10^{4.5} Lsun, which matches models of the post-merger evolution of super-Chandrasekhar mass white dwarfs. The high stellar temperature and the nebular size argue for a short remaining lifetime of the star, which will produce a bright optical and high-energy transient upon collapse. Our observations indicate that super-Chandrasekhar mass white dwarf mergers can indeed avoid a thermonuclear explosion as Type Ia supernova, and provide empirical evidence for magnetic field generation in stellar mergers.

Tables

Table 1. Table 1: Stellar parameters and surface abundances for WS35.

$\log \left(L_{\ast }/L_{\odot }\right)$	$4.51\pm 0.14$
$T_{\ast }$ (K)	$200,{000}_{-22,000}^{+12,000}$
$R_{\ast }$ ($R_{\odot }$)	${0.151}_{-0.013}^{+0.035}$
$\dot{M}$ ($M_{\odot }{\mathrm{yr}}^{-1}$)	($3.0\pm 0.4)\times 10{}^{-6}$
$D$	$10$
$v_{\infty }$ ($\mathrm{km}{s}^{-1}$)	$16,000\pm 1,000$
$v_D$ ($\mathrm{km}{s}^{-1}$)	$250$
$\beta$	$1.0$
$d$ (kpc)	${3.07}_{-0.28}^{+0.34}$
$E\left(B-V\right)$ (mag)	0.8
$R_V$	3.1
He (mass fraction)
C (mass fraction)	$0.2\pm 0.1$
O (mass fraction)	$0.8\pm 0.1$
Fe-group (mass fraction)	$1.6\times {10}^{-3}$

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, Universität Bonn, Auf dem Hügel 71, 53121 Bonn, Germany

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

Astronomický ústav, Akademie věd České Republiky, Fričova 298, 251 65 Ondřejov, Czech Republic

South African Astronomical Observatory, PO Box 9, 7935 Observatory, Cape Town, South Africa

Southern African Large Telescope Foundation, PO Box 9, 7935 Observatory, Cape Town, South Africa

Special Astrophysical Observatory of the Russian Academy of Sciences, Nizhnii Arkhyz, 369167, Russia

During our search for mid-infrared circumstellar nebulae[12, 13], we discovered a new object in the constellation Cassiopeia (Fig. 1) using data of the Wide-field Infrared Survey Explorer (WISE)[14]. At 22 $\mu$m the nebula (named WS35; cf. ref.[13]) appears as a circular shell with ragged edges and angular radius of $\approx 75$ arcsec. The higher contrast 22 $\mu$m image of WS35 shows a diffuse halo with a radius of $\approx 110$ arcsec surrounding the shell. For the distance to WS35 of $\approx 3$ kpc (see below), the linear radii of the shell and halo are, respectively, $\approx 1.1$ and 1.6 pc. The shell is also visible in the WISE 12 $\mu$m image, where it appears as a diffuse circular nebula of the same angular size as the 24 $\mu$m shell. Surprisingly, despite of the moderate extinction towards WS35 (see below) the infrared nebula has no optical counterpart in the INT Photometric H$\alpha$ Survey of the Northern Galactic Plane (IPHAS)[15] (see Fig. 1). We identified the central star of WS35 with an optical star ($V\approx$15.5 mag; see Methods) located at right ascension 00 h 53 m 11.21 s and declination +67*∘* 30′ 02.1*′′*. Below, we use the same name for the nebula and its central star.

Optical follow-up spectroscopy of WS35 with the Russian 6-m telescope (see Methods) revealed an emission-line dominated spectrum of the central star, reminiscent of oxygen-rich Wolf-Rayet (WO-type) stars (Fig. 2). However, WS35’s emission lines are stronger and broader than those of even the most extreme WO stars. Most notably, the O vi $\lambda\lambda$3811, 3834 emission doublet shows an equivalent width of EW(O vi)$\approx 2300$ Å  and a full width half measure of $\approx 300$ Å. We note that no nebular lines are visible in the obtained long-slit spectrum.

We analysed the optical spectrum of WS35 using the Potsdam Wolf-Rayet (PoWR) code for expanding stellar atmospheres (cf. ref.[16]), tailored for WS35 (Methods). In Fig. 2, we compare our best-fitting model with the observed spectrum. The line spectrum is reproduced well, except of two missing emission lines at 4340 and 6068 Å  which are likely due to line transitions of highly ionized neon (Ne viii; cf. ref. [17]), which are not included in the present modelling. Our model fit yields a stellar temperature at the base of the wind of $200,000^{+12,000}_{-22,000}$ K, and a chemical composition dominated by oxygen and carbon with mass fractions of $0.8\pm 0.1$ and $0.2\pm 0.1$, respectively. Remarkably, the absence of the He ii $\lambda$4686 emission line implies an essentially helium-free composition, with a conservative upper limit on the helium surface mass fraction of $<0.1$. From the width of the O vi $\lambda\lambda$3811, 3834 emission we derive a terminal wind velocity of $v_{\infty}=16,000\pm 1,000\,{\rm\,km\,s^{-1}}$. The strength of the emission lines is best reproduced with a mass-loss rate of $\dot{M}=(3.0\pm 0.4)\times 10^{-6}\rm\,M_{\odot}\,{\rm yr}^{-1}$ (where $\rm\,M_{\odot}$ is the solar mass) with a wind clumping factor of 10 (Methods).

Based on the distance of $3.07^{+0.34}_{-0.28}$ kpc derived from the Gaia data release 2 (ref. [10]), our model reproduces the observed flux distribution from ultraviolet to infrared with a luminosity of $\log(L/\rm\,L_{\odot})=4.51\pm 0.14$ (where $\rm\,L_{\odot}$ is the solar luminosity) and with an interstellar extinction of $E(B-V)=0.8$ mag (see Fig. 2 and Methods). This luminosity is five times lower than that of even the faintest massive WO-type Wolf-Rayet stars[18] ($\log(L/\rm\,L_{\odot})>5.2$), and at least three times higher than those of the known low-mass [WO]-type central stars of planetary nebulae[19] ($\log(L/\rm\,L_{\odot})<4.0$).

Very few hot stars free of hydrogen and helium are known in the Milky Way. Besides the rare DQ white dwarfs[20], these are two hot white dwarfs with unusually high masses ($0.7\dots 1.0\,\rm\,M_{\odot}$) which have been suggested to have formed through the merger of two carbon-oxygen white dwarfs[21]. Recent models for the evolution of super-Chandrasekhar mass carbon-oxygen white dwarf merger remnants by Schwab et al.[8] match the properties of WS35 remarkably well. Not only does their fiducial model fit the Hertzsprung-Russell diagram position of WS35 (Fig. 3). They also predict an extended episode of extreme mass loss during and shortly after the merger event in an expanded cool supergiant stage which will form a slowly-expanding hydrogen- and helium-free circumstellar nebula with an expansion velocity of about $100\,{\rm\,km\,s^{-1}}$. After the settling of the post-merger product in a hot and compact stage, its ultraviolet emission ionizes the surrounding nebula. Its current $\sim 100$ times faster wind sweeps the preceding slow material into a shell, which appears inside the density-bounded H ii  region, which forms the diffuse halo (see Fig. 1). The expected hydrogen- and helium-free composition of the nebula and the high temperature of its central star (leading to triple ionization of oxygen) suggest that no optical nebular lines form and that the observed mid-infrared emission from the nebula is dominated by the [O iv] 25.89 $\mu$m and [Ne v] 14.32 and 24.32 $\mu$m lines (cf. ref.[22]). This naturally explains why the nebula WS35 does neither appear in the IPHAS image nor in our long-slit spectrum.

The merging white dwarf scenario also addresses the major puzzle about WS35, namely the extreme width of its emission lines. A velocity of $16,000\,{\rm\,km\,s^{-1}}$ exceeds the stellar escape speed about 10 times, and is typical for supernovae but so far unheard of for radiation driven winds. Pure radiation driving is in fact excluded, since the wind’s kinetic energy flux exceeds the total radiative luminosity of the star by a factor of two (Methods). The extreme velocity can, however, be explained in the framework of rotating magnetic wind models. Poe et al.[23] find that a rigidly co-rotating magnetic field can enhance the speed and mass outflow rate of radiation driven winds by more than a factor of three, at the cost of the star’s rotational energy. We find that a co-rotation speed of $16,000\,{\rm\,km\,s^{-1}}$ at the Alfvén point in WS35, where the inertia force starts to dominate over the magnetic forces[24], requires an Alfvén radius of about 10 stellar radii ($\sim 1.5\,\rm\,R_{\odot}$), which is achieved with a magnetic field strength of $\sim 10^{8}$ G. Since the whole post-merger evolution is expected to last $\sim 20,000$ yr (ref.[8]), it is plausible that the corresponding magnetic torques have not yet spun down WS35.

The generation of a strong magnetic field is indeed expected in stellar mergers[25]. 3D magneto-hydrodynamical models of white dwarf mergers find a magnetization of the merger product of $2\times 10^{8}$ G (ref.[9]). This compares to the peak of the $B$-field distribution of magnetic white dwarfs of several $10^{7}$ G (ref.[26]). The observations that the mean mass of magnetic white dwarfs is significantly higher than that of non-magnetic ones, and that nearly none of the known magnetic white dwarfs have a companion star provide strong evidence for magnetic field generation by white dwarf merging[26].

Adopting the escape velocity from the expanded merger remnant of $\sim 100\,{\rm\,km\,s^{-1}}$ (cf. ref.[7]) and given the angular radius of the nebula of 1.6 pc, we derive an expansion age of $\sim 16,000$ yr for WS35. Together with the high temperature of the central star, this indicates that WS35 is close to the endpoint of its post-merger evolution. Since WS35 is more luminous than the 1.49 M*⊙* model of Schwab et al.[8], it appears likely that its mass also exceeds the Chandrasekhar limit, with the exciting perspective that it will produce a low-mass neutron star in the near future, accompanied by a high-energy transient and a fast evolving supernova[11]. It is also likely that the inner shell of the circumstellar nebula will persist until the explosion and, given the expected low mass of supernova ejecta, will affect the appearance of the supernova remnant $\sim 100$ yr after explosion.

The merging of two stars in a binary system is not a rare event. About 10 per cent of the massive main sequence stars are thought to be merger products[27], as well as a similar fraction of the known white dwarfs[28]. The very unusual wind of WS35 strongly supports the idea that stellar mergers can indeed produce highly magnetized stars, which would explain the magnetic stars of the upper main sequence[29] and the formation of the magnetic white dwarfs[26]. Our results also enlighten the ongoing debate on whether a super-Chandrasekhar mass merger of two carbon-oxygen white dwarfs leads to a Type Ia supernova. Here, WS35 appears to provide a counter example, which will likely produce a neutrino-flash and a gamma-ray burst[30], followed by a very fast and subluminous Type Ic supernova[11].