A Possible Surviving Companion of the SN Ia in the Galactic SNR G272.2-3.2

Type Ia supernovae (SNe Ia) are powerful calibrated candles, whose use as distance indicators in cosmology led to the discovery of the accelerated expansion of the universe (Riess et al. 1998; Perlmutter et al. 1999), and currently are major tools in the exploration of the nature of dark energy (Rose et al. 2020; Hayden et al. 2021). Besides, SNe Ia are the main producers of the Fe-peak elements in the universe (see, e.g., Branch & Wheeler 2017).

However, there is still a lack of knowledge concerning the exact nature of the progenitors of the SNe Ia, particularly their explosion mechanism and the kind of stellar systems from which they arise, both points being closely related. There is now a universal agreement that they are produced by the thermonuclear explosion of a white dwarf made of carbon and oxygen (a C+O WD), with a mass not far from the Chandrasekhar mass. But the explosion might be initiated close to the center of the star, when the mass reaches the Chandrasekhar limit owing to accretion of material from a close binary companion (Whelan & Iben 1973), or result from compression produced by the detonation of a helium layer close to the surface of the WD (Livne 1990; Livne & Arnett 1995) or by the collision with another WD (Rosswog et al. 2009). In the last two cases, the exploding WD would have a mass below the Chandrasekhar mass.

The WD progenitor of the SN Ia must be in a close binary system in all cases. The companion, the mass donor, may either be a star in any stage of thermonuclear burning (Whelan & Iben 1973; Nomoto 1982), called the single-degenerate (SD) scenario, or be another WD (Iben & Tutukov 1984; Webbink 1984), called the double-degenerate (DD) scenario. The core-degenerate (CD) model of SN Ia explosion (Kashi & Soker 2011; Soker 2013), in which a WD merges with the electron-degenerate core of an asymptotic giant branch (AGB) star, can be included within the DD scenario.

It is unknown what fractions of the observed SNe Ia correspond to each of the two scenarios. In the DD channel, in most cases considered, both WDs should be destroyed by the explosion, no bound remnant being left. There is a possible exception, though, in the case of explosions triggered by the detonation of a surface layer made of He, accreted by the exploding WD from a less massive WD companion: then, the outburst might happen when the mass donor has not yet been tidally disrupted. Due to its very high orbital velocity, the WD companion should be ejected as a hypervelocity star (v > 1000 km s⁻¹). This happens in the dynamically driven double-degenerate, double-detonation scenario (D⁶; Shen & Moore 2014; Shen & Schwab 2017; Shen et al.2018).

In the SD case, the binary companion of the exploding WD survives (Marietta et al. 2000; Pakmor et al. 2008; Pan et al. 2012a, 2012b). The surviving companion might be in any evolutionary stage—main sequence (MS), subgiant (SG), or red giant (RG)—and be a helium or an sdB star (see Wang & Han 2012; Maoz et al. 2014; Ruiz-Lapuente 2014, 2019, for reviews). Of course, the detection of such companions at the location of an SN Ia would confirm the SD scenario (for that particular SN at least and thus for some fraction of them).

In the hydrodynamical simulations of the impact of the SN Ia ejecta with the companion, different kinds of stars have been considered: MS stars (Marietta et al. 2000; Pakmor et al. 2012; Pan et al. 2012a; McCutcheon et al. 2022), SGs (Marietta et al. 2000; Pan et al. 2012a), helium stars (Pan et al. 2012a; Liu et al. 2013b), RG stars (Marietta et al. 2000; Pan et al. 2012a), and sdB stars (Bauer et al. 2019). These calculations predict the state of the companion just after the SN Ia explosion. Different amounts of mass have been stripped by the impact with the ejecta, and the stars are bloated and overheated. Those results provide the initial conditions to calculate the subsequent evolution of the companion stars.

The time evolution of possible SN Ia companions, on scales from hundreds to thousands of years, has been calculated by Podsiadlowski (2003) for an SG companion, by Pan et al. (2012b, 2014), Shappee et al. (2013), and Rau & Pan (2022) for MS and SG companions, by Pan et al. (2012b, 2014) for RGs, by Bauer et al. (2019) for sdB stars, and by Liu et al. (2022) for He stars. The calculations predict the changes in luminosity and effective temperature of the stars, starting from the time they recover hydrostatic equilibrium after experiencing the impact of the SN ejecta. The stars are then overluminous as compared with their previous state, and they evolve, on thermal timescales, to meet the characteristics corresponding to the new mass and thermonuclear burning stage. Pan et al. (2014) have also calculated the chemical pollution of the atmospheres by the ejecta.

One effect due to the star having been in a close binary system previous to the explosion are high space velocities (due to their orbital velocities prior to the disruption of the binary, in addition to the kick imparted by the ejecta). ¹⁰ Therefore, when searching for possible companions within the remnants of recent SNe Ia, one should look for high spatial velocities, paying attention to the past trajectories, and also for anomalous positions in the color–magnitude and color–color diagrams of the stars close to the center of the supernova remnant (SNR). We should also look for possible chemical enrichment, in Fe-peak elements namely. Given the current observational means, only remnants of SNe Ia that took place in our Galaxy or in the Large Magellanic Cloud (LMC) have been explored in the search for surviving companions, at present.

Up to 14 SNRs of the Ia type have been identified in the Galaxy, and 12 have been identified in the LMC (Ruiz-Lapuente 2019). Of the former, only three have been explored (corresponding to the "historical" SNe Ia): those of SN 1572, or Tycho Brahe's SN (Ruiz-Lapuente et al. 2004, 2019; González Hernández et al. 2009; Kerzendorf et al. 2009, 2013, 2018a; Bedin et al. 2014); SN 1604, or Kepler's SN (Kerzendorf et al. 2014; Ruiz-Lapuente et al. 2018); and SN 1006 (González Hernández et al. 2012; Kerzendorf et al. 2012, 2018b; Shields et al. 2022). No indisputable candidate companion has been found in any of them. In the case of SN 1006, the absence of candidates points to a DD origin of the SN. In the case of SN 1604, the same absence joined to the characteristics of the SNR suggested the CD scenario (González Hernández et al. 2012). A candidate has been found for SN 1572, but the identification is in dispute (see the references above).

Five SNRs of Ia type have been explored in the LMC: SNR 0509–67.5 (Schaefer & Pagnotta 2012; Litke et al. 2017), SNR 0519–69.0 (Edwards et al. 2012; Li et al. 2019), SNR N103B (Li et al. 2017), SNR 0505–67.9 (DEML71), and SNR 0548–70.4 (Li et al. 2019). No clear surviving companion candidate has been found in any of them, but a star in N103B has characteristics compatible with being a surviving SG (Li et al. 2017), and two other stars, in SNR 0519–69.0 and DEML71, respectively, have large radial velocities and might also be SN companions (Li et al. 2019).

Out of the still-unexplored SN Ia Galactic SNRs, most are at large distances and located close to the Galactic plane, which causes them to be very heavily reddened.

That is not the case, however, for SNR G272.2-3.2, at a distance of ∼1–3 kpc. The EDR3 of Gaia now provides the parallaxes, proper motions, and photometry, allowing a first exploration of the central region of this SNR. Knowledge of the parallaxes allows us to select the stars, close to the center of the SNR, that are at distances compatible with that of the remnant. We look for peculiar proper motions and compare the H-R diagram of the sampled stars with the evolutionary paths predicted for different types of surviving companions. That already allows us to exclude the presence of several kinds of proposed candidates and to select stars deserving further analysis.

In the next section we summarize the characteristics of the SNR G272.2-3.2 and define the search area for the possible companion star. Observations are described in Section 3. Proper motions, their transformation to tangential velocities, and a kinematic outlier are treated in Section 4. Reddening of the observed field is discussed, together with the stellar spectra obtained, in Section 5. The stellar parameters of our unique peculiar star are determined and a chemical analysis is performed in Section 6. In Section 7, the characteristics of this star are further discussed. Color–magnitude and H-R diagrams, along with their comparison with the evolutionary tracks predicted for different types of companions, are dealt with in Section 8. Exploration of more extended areas than in Section 2 and the search for the possible presence of hypervelocity stars are examined in Section 9. All results are summarized and conclusions drawn in Section 10.

The SNR G272.2-3.2 was discovered in X-rays by Greiner & Egger (1993) during the ROSAT All Sky Survey, with details given in Greiner et al. (1994). Radio observations by Duncan et al. (1997) measured a diameter of ∼15' for the remnant. It was later studied, in X-rays, by Harrus et al. (2001), Lopez et al. (2011), McEntaffer et al. (2013), Yamaguchi et al. (2014), and Kamitsukasa et al. (2016), with mounting evidence, from the measurement of overabundances of Ar, Ca, Si, S, Fe, and Ni, that it was produced in an SN Ia.

Chandra observations have provided measurements of chemical abundance ratios that are in good agreement with the predictions for delayed detonation models of SN Ia explosions (Sezer & Gök 2012).

The SNR is at a distance $d={1.8}_{-0.8}^{+1.4}$ kpc (Greiner et al. 1994) or d ∼ 2–2.5 kpc (Harrus et al. 2001; Kamitsukasa et al. 2016). Its age is estimated to be 7500${}_{-3300}^{+3800}$ yr (Leahy et al. 2020; Xiang & Jiang 2021).

It is located about 110 pc below the Galactic plane. The radius of the remnant is about 8', and its centroid lies at α_J200 = 09^h06^m457, δ_J2000 = −52^o07'03'' (Greiner & Egger1993), which corresponds to the Galactic coordinates l = 272°12'369, b = −3°10'344.

We have searched in the Gaia EDR3 database for the stars within a radius of 11' (thus extending beyond the whole SNR) and with parallaxes corresponding to distances 1 kpc ≤d ≤ 3 kpc. That has produced a sample of 3082 stars (see Figure 1). The 11' radius is slightly above the arc described by a star moving at 500 km s⁻¹, perpendicularly to the line of sight and a distance of 2 kpc, in 12,000 yr. Wider search radii are considered, and the corresponding results are presented and discussed in Section 9.

Zoom In Zoom Out Reset image size

Besides parallaxes and proper motions, we have also extracted the Gaia photometry of the stars in our sample (left panel of Figure 2). The photometry in Figure 2 is still uncorrected for interstellar extinction and reddening.

Zoom In Zoom Out Reset image size

3.1. Gaia EDR3

3.2. DECaPS

3.3. Spectra

Peculiar space velocities as compared with those of the surrounding stars are among the likely characteristics of surviving companions of SN Ia explosions. In Figure 3 we show the distribution of proper motions in R.A. (left) and decl. (right). The mean proper motion in R.A. (${\mu }_{\alpha }^{* }$) is −4.8 mas yr⁻¹, with a standard deviation σ = 3.12 mas yr⁻¹, while in decl. (μ_δ ) we have a mean of 4.44 mas yr⁻¹ with σ = 3.11 mas yr⁻¹. There is a star, Gaia EDR3 5323900211541075328 (MV-G272) (R.A. = 09^h 06^m 2466; decl. = −52° 03' 09684, G = 19.854 mag), with ${\mu }_{\alpha }^{* }$ = −22.80 mas yr⁻¹; μ_δ = 30.60 mas yr⁻¹, which is the only extreme outlier in the two proper-motion distributions: at 5.8σ from the mean in R.A. proper motion and at 8.4σ in decl.

Zoom In Zoom Out Reset image size

The distribution of total proper motions ($\mu \,=\sqrt{(}{({\mu }_{\alpha }^{* })}^{2}\,+{({\mu }_{\delta })}^{2})$) is shown in Figure 4 (left panel). Since we know the distances to the stars, we can calculate their total velocities perpendicularly to the line of sight, v_tan (using the expression ${v}_{\tan }=4.7485\times \mu /\varpi$, with ϖ being the parallax). The resulting distribution is shown in the right panel of Figure 4. Here we see again the same outlier, with a total proper motion of 38.15 mas yr⁻¹ (which is 8.9σ above the mean). The distance to the star, from its parallax, is $d={1.32}_{-0.39}^{+1.00}$ kpc, which gives a tangential velocity ${v}_{\tan }={239}_{-70}^{+181}$ km s⁻¹ (5.4σ above the mean).

Zoom In Zoom Out Reset image size

Total speeds of that order are expected for MS or SG companions of an SN Ia (Han 2008; see also Table 1 in Pan et al. 2014), so this star deserves further study.

Since the SNR G272.2-3.2 has a minimum age of ∼4500 yr, any possible surviving companion should have traveled an appreciable distance from the site of the explosion by now. From the proper motions measured, we can infer its position at the time of the SN outburst. In Figure 5, that is made for an age of 8000 yr. We see that star MV-G272, located at the periphery of the SNR at present, was very close to the center by then.

Zoom In Zoom Out Reset image size

The Gaia EDR3 does not provide information on the radial velocities v_r of any of those stars. From a spectrum of MV-G272 (see the next section), we have measured a barycentric velocity v_bar = 92.6 ± 0.5 km s⁻¹ (v_r = 77.3 km km s⁻¹ in the LSR). That gives ${v}_{\mathrm{tot}}={256}_{-70}^{+181}$ km s⁻¹ (barycentric) for this star.

It can be seen, from Figures 3 and 4, that there are a few stars at σ > 3 in the corresponding distributions, but, in view of their positions relative to the SNR and of their trajectories, they are not viable candidates to be companions of the SN.

4.1. The Kinematics of Star Gaia EDR3 5323900211541075328/MV-G272

The star MV-G272 looks like a possible candidate to be the surviving companion of the SN Ia that gave rise to SNR G272.2-3.2. Its kinematics is analyzed in more detail below. We will look at the motion in Galactic coordinates, taking for comparison the current Besançon model of the Galaxy. ¹³ Since the star is an M1–M2 dwarf with solar metallicity (see the next section), we will only use, from the model, M dwarfs with the same metallicity and located at distances 1 kpc ≤ d ≤ 2 kpc, like the candidate star. The resulting distributions in ${\mu }_{l}^{* }$, μ_b , and v_r of the model stars are shown in Figure 6.

Zoom In Zoom Out Reset image size

With ${\mu }_{l}^{* }$ = −37.96 mas yr⁻¹, star MV-G272 is at 6.1σ above the mean, while its μ_b = 3.85 mas yr⁻¹ is at 1.4σ only, and v_r = 77.1 km s⁻¹ in the LSR is at 2.4σ. Thus, only the motion along the Galactic plane is really peculiar as compared with the Besançon model.

We can also select the M-dwarf population within our sample, equally at distances 1 kpc ≤ d ≤ 2 kpc, and look at its distribution in μ_l and μ_b . In this way we know how star MV-G272 moves as compared with the surrounding stars of the same type. The distributions are shown in Figure 7. This star is a clear outlier in ${\mu }_{l}^{* }$, at 7.4σ above the mean, while it is only at 1.7σ in μ_b .

Zoom In Zoom Out Reset image size

To investigate further the kinematics of star MV-G272, we have also calculated its orbit, using GravPot16 (Fernández-Trincado 2017). In Figure 8 we show the 3D orbit and its projections on the X–Y, X–Z, and Y–Z planes of the referential system of the Galaxy. We see that the projection of the orbit on the X–Y plane is very eccentric, with e = 0.447. Only 13 stars in our sample have radial velocities determined in Gaia EDR3, and only these can thus be used to calculate eccentricities. The mean eccentricity is 〈e〉 = 0.167, with σ = 0.063. Therefore, that of star MV-G272 is 4.5σ above the mean, as illustrated in Figure 9.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Measured Gaia magnitudes are affected by absorption due to intervening gas and dust. In order to compare the observed photometry with theoretical models of the evolution of possible stellar companions after the SN Ia explosion, the extinction in different bands must be accurately estimated and subtracted.

There are only four stars in our sample that have their parameters T_eff, log g, and [Fe/H] determined in the Gaia EDR2 (see Table 2).

Using the expressions by Carrasco, ¹⁴ we translate their magnitudes to the Johnson's UBVRI system and then compare the measured colors with those corresponding to stars with the same parameters (Houdashelt et al. 2000). From the color excesses, we deduce the extinctions A_V , A_R , A_I for each of the four stars. Since the stellar parameters from Gaia EDR2 do not seem very accurate, we do not expect a perfect coincidence in the values obtained for the different stars, but the average value for A_V is 1.65 mag, in agreement with that obtained from the spectral fits below.

Another estimate comes from fitting spectra of star MV-G272, taken with the MIKE spectrograph at the 6.5 m Clay telescope and with the Goodman spectrograph at the 4.1 m SOAR telescope (both covering similar wavelength ranges). We have used the PyHammer tool (Kesseli et al. 2017, 2020) to infer that the best-fit spectral type is M1V−M2V. This code uses a set of templates for different spectral types and luminosity classes with a discrete set of metallicity values with a step of ∼0.5 dex, created from observed SDSS/BOSS spectra at R ∼ 2000. In Figure 10 we show the fits of both spectra to the template for an M1V star, with solar metallicity and a reddening of E(B − V) = 0.532 (top panels), and to an M2V star, with negligible reddening and solar metallicity too (bottom panels). Good fits to an M1V spectrum can also be obtained for smaller reddening and a somewhat higher metallicity ([Fe/H] = +0.5). We prefer the first fits, which correspond, for A_V = 3.1 × E(B − V), to A_V = 1.65 mag and coincide with the estimate made from photometry.

Zoom In Zoom Out Reset image size

As mentioned above, we have measured a radial velocity of v_r = 92.6 ± 0.5 km s⁻¹ (barycentric) or v_r = 77.3 ± 0.5 km s⁻¹ (LSR), using the MIKE spectrum.

A deeper and more complete analysis of the MIKE spectrum is made in the next section.

We have analyzed the high-resolution MIKE spectrum (R ∼ 28,000) to try to estimate global metallicity and some element abundances, from individual lines available in the red part of the MIKE (Bernstein et al. 2003) spectrum. The star is quite faint, which gives an estimated S/N ∼ 18 at 7500 Å. The spectrum has a total exposure time of 13,200 s. For comparison, we also analyzed two high-resolution CARMENES VIS spectra (Reiners et al. 2018) of two stars classified as M1V stars and the solar ATLAS spectrum (Kurucz et al. 1984) as a reference (see Appendix and discussion below). The two CARMENES spectra of these stars have been recently analyzed (Marfil et al. 2021) with the SteParSyn code (Tabernero et al. 2021, 2022), providing the following set of parameters: ${T}_{\mathrm{eff}}/\mathrm{log}g/$[Fe/H] = 3603/4.99/−0.52 for star Karmn J00183+440 (GX And) and ${T}_{\mathrm{eff}}/\mathrm{log}g/$[Fe/H] = 3825/4.94/−0.04 for star Karmn J05415+534 (HD 233153). Deriving the metallicities of M dwarfs even with high-quality spectra at high resolution is a challenging exercise, with differences of 0.3 dex from different methods (Passegger et al. 2022). These analyses have been done from individually resolved lines at very high resolution of these very high quality spectra, using those codes mentioned above.

We have also analyzed single CARMENES VIS spectra of these two stars and the solar ATLAS spectrum, all degraded to a resolving power of 28,000 and with injected noise to S/N ∼ 18 at 7500 Å, to match the MIKE spectrum resolution and quality. Thus, we implemented a Bayesian Python code that compares the observed spectrum with a synthetic spectrum in the spectral range 7000–8750 Å. Both the observed spectra and the synthetic spectra are normalized using a running mean filter with a width of 200 pixels at a dispersion of 0.069 Å pixel⁻¹ (see figures in Appendix). We performed a Markov Chain Monte Carlo with 5000 chains implemented in emcee (see Foreman-Mackey et al. 2013), sufficient to get a statistically significant result. We use a small 3 × 3 × 3 grid of synthetic spectra with values ${T}_{\mathrm{eff}}/\mathrm{log}g/$[Fe/H] of 3500–4500/3.0−5.0/−1.0 to 0.5 and steps of 250 K/0.5 dex/0.5 dex, computed with the SYNPLE ¹⁵ code, assuming a microturbulence ξ_mic = 0.8 km s⁻¹, and ATLAS9 model atmospheres with solar α-element abundances ([α/Fe] = 0; see Castelli & Kurucz 2003). Details of the fits are shown in Appendix.

The model includes as free parameters the effective temperature T_eff, surface gravity $\mathrm{log}g$, metallicity [Fe/H], rotational velocity V_rot, and relative radial velocity V_rel (Figure 11). We ran a simulation leaving free all the parameters and found that the simulation converges to a T_eff value at the lower edge of the grid. A similar result with a lower T_eff value by about 250 K than those obtained at high resolution was obtained when analyzing the two CARMENES spectra. Thus, we decided to run a simulation by fixing the T_eff at 3800 K for our target star. The posterior distributions of the simulation are displayed in Figure 11, which provide the values $\mathrm{log}g/$[Fe/H] = 4.46/−0.32. For the two degraded CARMENES spectra we got ${T}_{\mathrm{eff}}/\mathrm{log}g/$[Fe/H] = 3580/4.37/−0.69 for star Karmn J00183+440 (GX And) and ${T}_{\mathrm{eff}}/\mathrm{log}g/$[Fe/H] = 3805/4.67/−0.04 for star Karmn J05415+534 (HD 233153). As seen in Figure 12, the rotational velocity is not resolved with the instrumental FWHM of 10.7 km s⁻¹, providing a value consistent with zero and an upper limit at 3σ of V_rot < 3 km s⁻¹. The heliocentric radial velocity of the star is estimated at 92.60 ± 0.5 km s⁻¹, which translates into V_LSR = 77.3 ±0.5 km s⁻¹ in the LSR.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We analyzed the solar ATLAS spectra to get the reference solar element abundances, including all the relatively isolated features available in the MIKE spectrum of our target star, including five Fe lines, seven Ti lines, two Cr and Na lines, one Ni line, and one Al line. For that exercise we used a grid of models with values ${T}_{\mathrm{eff}}/\mathrm{log}g/$[Fe/H] of 5500–6000/4.0–5.0/−0.5 to 0.5 and steps of 250 K/0.5 dex/0.5 dex, assuming a fixed ξ_mic = 0.95 km s⁻¹. The analysis of the degraded solar spectrum provided all the element abundances within [X/H] = −0.13 for Ti and −0.03 dex for Na. The Ca features were discarded because they provided very different results for Ca i and Ca ii lines even in the solar case. The tentative abundance ratios found in the target star are [Na/H] = −0.10, [Al/H] = −0.23, [Ti/H] = −0.05, [Cr/H] = −0.08, and [Ni/H] = −0.23. These values may indicate a slight enhancement at 0.1–0.2 dex in all the element abundances with respect to the metallicity value of [Fe/H] = −0.32. We consider them very tentative with this methodology in M dwarfs given also the S/N of the MIKE spectrum. In any case, with these stars being almost fully convective, any captured material from the SN ejecta should be strongly diluted.

The characteristics of star MV-G272 are summarized in Table 3.

We know that, at present, star MV-G272 has the characteristics of an M1–M2 dwarf. Assuming that it was the companion of the SN Ia that gave rise to the SNR G272.2-3.2 and that its mass and radius at the time of the explosion were similar to the present ones, it should have been in close orbit with a 1.4 M_⊙ WD and filling its Roche lobe. We can calculate at which velocity it should have been ejected.

We would have, for the orbital motion of the star around the center of mass of the binary,

$$\begin{eqnarray}&&{v}_{\mathrm{orb}}^{2}\,=\,\displaystyle \frac{{{GM}}_{\mathrm{WD}}}{a(1+q)},\end{eqnarray} \tag{ 1 }$$

where a is the orbital separation and q ≡ M_comp/M_WD. In our case we have M_WD = 1.4M_⊙ and M_comp = 0.44M_⊙, corresponding to an M1 dwarf (Pecaut & Mamajek 2013). Hence, in our case q = 0.314. The radius of an M1 dwarf is R_M = 0.446 R_⊙ (same source).

We have, on the other hand, the Eggleton (1983) approximate formula for the Roche lobe radius R_L of the secondary star in a binary:

$$\begin{eqnarray}&&{R}_{L}=a\left[\displaystyle \frac{0.49}{0.6{q}^{-2/3}+\mathrm{ln}(1+{q}^{1/3})}\right].\end{eqnarray} \tag{ 2 }$$

We thus have for the orbital velocity, by making R_L = R_M in Equation (2) to obtain a and substituting it in Equation (1),

$$\begin{eqnarray}&&{v}_{\mathrm{orbit}}^{2}\,=\,\displaystyle \frac{0.49\ {{GM}}_{\mathrm{WD}}}{{R}_{M}(1+q)[0.6{q}^{-2/3}+\mathrm{ln}(1+{q}^{1/3})]},\end{eqnarray} \tag{ 3 }$$

and then, rounding to unity, v_orbit = 350 km s⁻¹. We have measured a total velocity ${v}_{\mathrm{tot}}={256}_{-70}^{+181}$ km s⁻¹ for star MV-G272 (vector sum of tangential and radial velocities).

The orbital velocity is just an upper limit to the actual one before explosion, since some mass should have been stripped by the impact of the SN ejecta and the pre-explosion radius and orbital separation should also have been larger then.

Concerning the rotational velocity, even before explosion rotation might have been slowed down owing to angular momentum loss from the mass transfer to the WD. The collision with the ejecta of the SN can drastically reduce the rotational velocity. This has been shown by Liu et al. (2013a) and Pan et al. (2014). In the 3D hydrodynamical simulations of Liu et al. (2013a), the rotational velocity of the companion is reduced to only 14%–32% of its pre-explosion value. Similar results are quoted by Pan et al. (2014), with references to their previous work (Pan et al. 2012b, 2013). An extra mechanism to slow down the rotation of the companion star after the impact of the SN ejecta would act during the evaporation phase of the surface layers of the star (those that have not been ablated by the impact but have absorbed enough energy to become unbound). If the wind remains tied to the surface of the star by the magnetic field and is only lost at significant distances above the surface, it will carry a lot of angular momentum, thus significantly reducing the rotational velocity (this idea is being explored by X. Meng et al. 2023, in preparation).

Finally, it must be remembered that what is actually measured is v_rot sin i, where i is the angle made by the rotation axis with the line of sight.

Pan et al. (2012a) have calculated the amount of contamination by Fe and Ni of the surfaces of SN Ia companions. They obtained ∼10⁻⁵ M_⊙ for MS star companions, ∼10⁻⁴ M_⊙ for He star companions, and ∼10⁻⁸ M_⊙ for RG companions (see also Pan et al. 2014). The observed contamination would, however, depend on the degree of dilution of the contaminants with the stellar envelope. Even in an early M dwarf, most of the mass, from the surface down to close to the central layers, is convective. We should thus expect a strong dilution of the material captured from the ejecta, and thus only moderate overabundances of Fe-peak elements.

From the Gaia photometry and parallaxes, we can now construct H-R diagrams for our sample of stars. They can then be compared with model predictions for different types of possible survivors from SN Ia explosions. There are only a few theoretical calculations of the evolution of SN Ia companions after being hit by the SN ejecta. Podsiadlowski (2003) modeled the evolution of an SG star of 2.1 M_⊙ for up to 10,000 yr after the explosion, and later Shappee et al. (2013) did the same for an MS companion of 1 M_⊙. In both cases, mass stripping from the impact was modeled as a fast wind, and energy injection into the layers of the companion that remained bound was parameterized. The results, for the luminosities and effective temperatures of the companions, 10,000 yr after the explosion, very widely differ, as can be seen by comparing Figure 1 in Podsiadlowski (2003) with Figure 4 in Shappee et al. (2013).

Di Stefano et al. (2011; see also Justham 2011) have calculated the evolution of SN Ia companions for the case in which there is long enough delay between the end of mass transfer and the explosion (due to the fast rotation of the WD) to allow the companion to become a second WD before the explosion takes place. Their calculations, however, stop at this point, and there are no existing hydrodynamical simulations of the collision of the SN material with the WD, a prerequisite to know its state, thousands of years after the explosion.

Meng & Li (2019; see also Meng & Luo 2021) have calculated the luminosities and colors of possible MS companions ending as subdwarf (sdB) stars at the time of the SN explosion, but, again, neither the effects of the impact on them nor their subsequent evolution are included there. Bauer et al. (2019) have modeled the collision of SN Ia ejecta with an sdB star and also the evolution after the impact, providing various useful observational predictions. Pan et al. (2014), on the other hand, start from 3D hydrodynamic models of the companions after being hit by the SN ejecta and follow their evolution hydrodynamically until hydrostatic (but not thermal) equilibrium is recovered. Those 3D models are then projected into 1D models, whose subsequent evolution is calculated using the MESA (Modules for Experiments in Stellar Astrophysics) code. ¹⁶ In this way they predict the evolutionary tracks for MS companions up to 9000 yr after being hit by the SN ejecta and those for He star companions until 1000 yr after the explosion.

In Figure 13 we show the tracks followed by MS and He star companions with several masses and pre-explosion orbital parameters, taken from Pan et al. (2014). Model characteristics of MS companions are given in Table 4 (which reproduces part of their Table 1). The tracks are plotted on the g versus g − r plane (Sloan colors), assuming a distance of 2 kpc. The positions of the sdB stars from Meng & Li (2019) are equally shown. Comparison is made with the stars in our Gaia sample and also with the larger sample from DECaPS, which includes stars within a wider range of distances. Both samples are corrected for reddening (see Section 5).

Zoom In Zoom Out Reset image size

Table 3. Characteristics of Star Gaia EDR3 5323900211541075328 (MV-G272)

Parameter	Value
${\mu }_{\alpha }^{* }$ (mas yr−1)	−22.79
μδ (mas yr−1)	30.60
μ (mas yr−1)	38.15
${\mu }_{l}^{* }$ (mas yr−1)	−37.96
μb (mas yr−1)	3.85
d (kpc)	${1.32}_{-0.39}^{+1.00}$
${v}_{\tan }$ (km s−1)	${239}_{-70}^{+181}$
vr (km s−1)	77.3 ± 0.5 (LSR)
vr (km s−1)	92.6 ± 0.5 (barycentric)
vtot (km s−1)	${256}_{-70}^{+181}$
G mag	19.85
GBP mag	21.03
GRP mag	18.77
Spectral type	M1–M2
Luminosity class	V
[Fe/H]	−0.32 ± 0.04
M (M⊙)	0.44-0.50
R (R⊙)	0.446–0.482
Teff (K)	3600–3850
log g	${4.46}_{-0.11}^{+0.10}$
log(L/L⊙)	−1.54/−1.39

Download table as:  ASCIITypeset image

We clearly see that no sdB stars are present in any of the two samples. As for the He star companions, although a few stars from DECaPS lie not far from the ends of the tracks, this is consistent with the dispersion owing to the lack of distance boundaries there. MS companions are not close to our Gaia sample either. They will be discussed next.

In Figure 14 we compare, in the H-R diagram log(L/L_⊙) versus log T_eff, the evolutionary tracks for the MS models A–G in Table 3 with our Gaia sample.

Zoom In Zoom Out Reset image size

The tracks calculated by Pan et al. (2014), as mentioned above, follow the evolution of the possible companions up to 9000 yr after the explosion. That covers, at least in part, the estimated range of ages for the SNR. Only very few stars lie close to the ends of the evolutionary tracks A–G, and they do not display any kinematical peculiarity.

Star MV-G272 lies quite apart from the tracks, but its mass should be M ≃ 0.44–0.50 M_⊙, less than half the smaller mass in Table 3 and Figures 13 and 14 (that of model G). No simulations of the impact of the SN ejecta or calculations of the subsequent evolution are available for stars of such a small mass.

Very recently, Rau & Pan (2022) have extended the calculations of the post-impact evolution of MS companions down to 0.8 M_⊙, the result depending on the ratio of the orbital separation to the radius of the star at the time of the explosion, for a fixed explosion energy. They follow the evolution up to more than 10⁵ yr after the explosion. They find that, for the lowest mass considered (0.8 M_⊙), the luminosity remains constant and can be as low as ∼2L_⊙ from shortly after the explosion until ∼10³ yr later. From this point on, the luminosity starts decreasing. A constant luminosity stage is common to all the masses studied (0.8, 1, 1.5, and 2 M_⊙), with such luminosity decreasing fast when going from the higher to the lower masses (see their Figure 2). These new results appear consistent with an origin of star MV-G272 from the impact of SN Ia ejecta on an MS companion of low mass.

The simplest hypothesis for the origin would be that MV-G272 was a more massive MS star at the beginning of mass transfer to the WD. The mass was reduced, by the process of mass transfer to the WD, to a value close to the present one. There is little mass stripping and energy input in the explosion, due to the compactness of the star. Another possibility might be that MV-G272 was much more massive than MV-G272 even at the time of the explosion, a sizable fraction of its mass having been stripped by the impact of the ejecta. This is what happens, in variable extent, to the model stars in Table 3, although none of them end up having a mass as low as that of an M1–M2 dwarf, since they are initially too massive for that. A third possibility is that MV-G272 was an M dwarf already at the start of mass transfer. The WD should then have been quite massive. M dwarfs have never been favorites as possible SN Ia companions. They would transfer H-rich material to the mass-accreting WD at a slow rate, hence giving rise to nova-like outbursts, with these expelling most of the accreted mass. Wheeler (2012), however, has presented a model in which the combined magnetic fields of the WD and of the M star lock them together. A kind of magnetic bottle would then form and channel the mass transfer, so the WD would be accreting matter through a limited polar area. Accretion rates would be enhanced owing to the luminosity of that hot spot acting on the M-dwarf end of the bottle, mixing being inhibited by the magnetic field, and the accreted material being kept hot, thus avoiding thermonuclear runaway outbursts. Not being spun up by the accretion, the WD would be slowly rotating. Given the high numbers of M dwarfs, this mechanism might contribute to the observed SN Ia rates. A fourth and last possibility is, of course, that star MV-G272 was unrelated to the SNR G272.2-3.2, its high velocity being due to past interactions with other stars, but then its path within the remnant would be a most extraordinary coincidence. We have examined several possibilities. Explanations for high-velocity stars include disruption of a close binary by the supermassive black hole at the center of the Galaxy, with capture of a member and ejection of the other; similar three-body interaction involving a black hole at the center of a globular cluster; tidal shredding of dwarf galaxies; or ejection from a nearby galaxy. None of these mechanisms are likely to impart high velocity along the Galactic plane. Dynamical interaction between groups of massive stars leading to binary disruption has also been proposed, and that would happen in the disk, but no low-mass stars like MV-G272 would be ejected. There is no stellar stream toward the site of the SNR. There is no pattern of ejected stars from a globular cluster inside the SNR, and its position is far away from the supermassive black hole at the center of the Galaxy.

Although our initial exploration of the SNR G272.2-3.2 has produced a good candidate to be the surviving companion of the exploding WD that produced the remnant, one may wonder whether a more extended search would produce some other candidate, moving faster than MV-G272, as might be the case of some He star companions or of hypervelocity stars (produced by the D⁶ mechanism mentioned in the introduction). We will thus further explore the region around G272.2-3.2, in order to check for such extra possibilities.

In Section 2 and thereafter, a search radius of 11' around the centroid of SNR G272.2-3.2 has been adopted for the exploration. We remind that this radius corresponds to the angular distance covered by a star at 2 kpc from us, moving at 500 km s⁻¹ for 12,000 yr, perpendicularly to the line of sight. One might now take the lower limit to the distance to G271.2-3.2 (1 kpc) and the upper limit to its age (12,000 yr) and include stars with tangential velocities of up to 1000 km s⁻¹ (possible He star companions); a search radius of 422 results (almost four times that used for our initial search). We have also made such an extended exploration. In the left panel of Figure 15 we show the present and past positions (8000 yr ago) of the nine stars with total proper motions larger than that of star MV-G272, in the new area. We clearly see, from the figure, that none of the nine stars (whose characteristics are listed in Table 5) can have anything to do with the SNR. Stars with smaller proper motions (but still at more than 3σ above the average), located within the new ring, are not shown in Figure 15, but it is obvious that, with them being located outside the 11' radius around the centroid at present and moving more slowly across the sky than MV-G272, they can hardly even have been inside the area now covered by the remnant (which has an approximate radius of 9' only) 8000 yr ago.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 5. The Nine Stars with Total Proper Motions Larger Than That of MV-G272, within the 422 Radius from the Centroid of G272.2-3.2

Star	R.A.	Decl.	d	${\mu }_{\alpha }^{* }$	μδ	vtan
	(deg)	(deg)	(kpc)	(mas yr−1)	(mas yr−1)	(km s−1)
A	137.017	−52.146	${2.11}_{-1.31}^{+* }$	−25.641	76.381	807${}_{-501}^{+* }$
B	136.109	−52.488	${1.30}_{-0.32}^{+0.65}$	−39.558	30.548	308${}_{-75}^{+155}$
C	136.472	−52.608	${1.00}_{-0.20}^{+0.36}$	−35.620	34.783	236${}_{-47}^{+85}$
D	137.720	−51.903	${2.96}_{-1.94}^{+* }$	−26.766	35.898	629${}_{-412}^{+* }$
E	136.439	−52.648	${1.87}_{-0.61}^{+1.76}$	−18.969	37.134	370${}_{-120}^{+349}$
F	137.242	−51.755	${1.19}_{-0.56}^{+11.48}$	22.633	−34.213	232${}_{-109}^{+2236}$
G	137.333	−52.060	${1.43}_{-0.55}^{+2.40}$	−33.587	22.790	276${}_{-106}^{+462}$
H	137.392	−52.507	${1.03}_{-0.38}^{+1.54}$	35.488	−19.089	197${}_{-73}^{+294}$
I	136.212	−52.414	${2.57}_{-1.83}^{+* }$	−32.943	22.413	486${}_{-346}^{+* }$

Download table as:  ASCIITypeset image

Concerning the high-proper-motion stars in the extended 422-radius area, it must be noted that the sample comprises 54,035 stars. An estimated fraction of 0.002 of the stars in the solar neighborhood belong to the halo population (see, e.g., Konishi et al. 2015). Thus, the sample should include ∼100 of them, with velocities of at least 220 km s⁻¹ relative to the local standard of rest (Du et al. 2018). From its metallicity we know that MV-G272 is not one of these.

A possible scenario for the production of SNe Ia is, as we said in the introduction, the dynamically driven double-degenerate, double-detonation scenario (D⁶; Shen & Moore 2014; Shen & Schwab 2017; Shen et al. 2018). In this scenario, detonation of a He layer at the surface of the more massive, mass-accreting WD induces, by compression, its whole detonation. That would happen when not much mass has yet been transferred from the less massive WD, which has not been tidally disrupted. It would therefore be ejected from the system at its orbital velocity. Given that the system is extremely compact, the velocity should be very high (>1000 km s⁻¹). In fact, three objects have been found, in the Gaia EDR2, moving at 1000–3000 km s⁻¹, which might be former WD companions in a pre-SN system (Shen et al. 2018).

Since SNR G272.2-3.2 is comparatively young, even a possible hypervelocity former companion of the SN cannot have traveled very far. Even moving at 3000 km s⁻¹, perpendicularly to the line of sight, for 12,000 yr, its trajectory on the sky, assuming a distance of 2 kpc, would only reach 087 away from the site of the explosion. Thus, exploring the region up to a full degree from the centroid of the SNR is enough to catch any possible hypervelocity object produced by the explosion.

This new exploration has produced a sample of 112,704 stars, none of them with a total proper motion larger than that of star A in Table 5 and Figure 15. All the new high-velocity stars are outside the previously explored 11' radius, so they can in no way have originated from the SNR.

To be exhaustive, one can take a distance of 1 kpc, with the same tangential velocity of 3000 km s⁻¹. That gives a radius of 21. The sample then comprises 514,072 stars, and again there is no star with total proper motion surpassing that of star A.

From a Goodman spectrum, star A (Gaia EDR3 5323871314998012928) is an MV1−M2V star, with solar metallicity. However, due to the direction of its motion (see Figure 15), it cannot have originated from inside the SNR G272.2-3.2; its tangential velocity, if it were at a distance of 2.11 kpc (see Table 5), would be ∼807 km s⁻¹ (with a large error propagated from that on the parallax). From a MIKE spectrum (see Figure 16) we have measured its radial velocity, which is only of v_helio = 57.5 ± 0.4 km s⁻¹ or v_LSR = 42.2 ±0.4 km s⁻¹. We thus think that star A is at a distance close to its lower limit in Table 5, which places its total velocity well below the range of the hypervelocity stars.

A relevant question is whether hypervelocity stars like those identified by Shen et al. (2018) would be detectable at the distance of the SNR G272.2-3.2. Two of them, labeled D6-1 and D6-3 in their Table 1, are at distances of 2.1–2.3 kpc and have Gaia magnitudes G of 17.4–18.3, so they would have clearly been seen in our survey. As for D6-2 (LP 398-9), it is at a distance d = 0.84 ± 0.04 kpc only (from this last reference) and has G = 16.97 mag. Thus, relocated at 2 kpc (the mean distance of our survey), it would still have G = 18.85 mag and also be detected. In addition, it has been associated with a ∼10⁵ yr old SNR. And the evolution of a WD after being impacted, heated, and bloated by the SN ejecta should be cooling, contracting, and fading, so at only 12,000 yr (the upper limit for the age of G272.2-3.2) such an object should be more luminous than D6-2/LP 398-9, if anything.

Therefore, we can quite confidently conclude that no hypervelocity star of the type D⁶ (Shen et al. 2018) has been produced by the explosion giving rise to G272.2-3.2.

G272.2-3.2 is the SNR of a relatively recent (∼7500 yr old) SN Ia that had been unexplored up to now in the search of possible surviving companions of the SN, though being, by its distance and location in the Galaxy, accessible to observations at all wavelengths.

We have first used the parallaxes, proper motions, and photometry from the Gaia EDR3, to explore the region within a circle of 11' radius around the centroid of the SNR and within a distance 1 kpc ≤ d ≤ 3 kpc. That produced a sample of 3082 stars. The surveyed area is larger than the SNR and encloses it. We then looked for kinematical signatures of a possible SN Ia companion star. We also had the Gaia photometry, used in a subsequent step. We checked this photometry against that from DECaPS and found complete agreement between them.

From the statistics of the proper motions of the stars in our sample, one of them, MV-G272, appears as a clear outlier, with a total proper motion 8.9σ above the mean. We checked this peculiarity against the Besançon model of the Galaxy, which confirmed it. The peculiar motion is mostly along the Galactic plane. Spectra obtained for this star have allowed us to measure its radial velocity as well. The total velocity is ${v}_{\mathrm{tot}}={256}_{-70}^{+181}$ km s⁻¹, which falls within the range of velocities expected for low-mass companions of SNe Ia. Reconstruction, from the proper motions, of the past trajectory shows that the star, which is now near the periphery of the SNR, was at the center 6000–8000 yr ago. Given the long path traced by the star, this coincidence is most significant. Such a trajectory is unique among the 3082 stars of the sample.

Spectra obtained with the MIKE spectrograph at the 6.5 m Clay telescope and with the Goodman spectrograph at the 4.1 m SOAR telescope allowed the classification of MV-G272 as an M1–M2 dwarf, with solar metallicity, by comparing them with BOSS templates. They also showed that the extinction, in the direction of the SNR, was small, which is relevant for the photometry. The star thus has a mass M = 0.44–0.50 M_⊙ and radius R = 0.446–0.501 R_⊙. There is agreement between the measured total velocity of MV-G272 and the ejection velocity of an M1–M2 dwarf in close orbit with a 1.4 M_⊙ WD, when the binary is disrupted by an SN Ia explosion.

The spectrum obtained with the MIKE spectrograph at the 6.5 m Clay telescope has allowed us to establish the values of the stellar parameters of MV-G272 and also to make a chemical analysis of its surface through synthetic spectra analysis (see Section 6). We have then T_eff = 3800 K (the range can be 3600–3850 K, given the systematic uncertainties), log g = 4.46, and metallicity about solar: [Fe/H] = −0.32 (with 0.3 dex in error). We also looked for signs of overabundances of Fe-peak elements coming from the SN ejecta, but systematic uncertainties preclude any conclusion. In any case, with an M1–M2 dwarf being almost fully convective, a strong dilution of this material should be expected.

The Gaia EDR3 photometry, together with the estimate of the reddening and the knowledge of the distances, allows us to construct the color–magnitude and H-R diagrams of the stars of our sample. They are compared with the existing models of the evolution of proposed SN Ia companions. That has allowed us to discard the presence of RGs, He stars, and sdB companion stars in our sample.

The models for MS companions are more luminous and hotter than the sampled stars. MV-G272 is fainter and cooler than any of the model stars, but all the models are for much more massive stars, so the comparison is not significant.

To be exhaustive, the possibility that the SNe were produced through the D⁶ mechanism has been checked by the exploration of a circle with a 21 radius around the SNR. None has been found.

We have examined all evolutionary paths that might have led to the SN Ia that produced SNR G272.2-3.2.

We have a very kinematically peculiar star, MV-G272, with all signs of having been ejected by an explosion taking place at the center of an SN Ia SNR. Such characteristics make it unique. Although its being a late-type, small star could come as a surprise, the evidence in its favor is very solid, coming from its kinematics and its trajectory inside the SNR.

In conclusion, we have found a possible companion star of the SN Ia that resulted in SNR G272.2-3.2, with much evidence in its favor. This would be the case, therefore, of an SD scenario involving an M-dwarf star. Since, from the chemical abundances in its ejecta, the explosion that gave rise to SNR G272.2-3.2 was a normal SN Ia and M dwarfs are the most abundant stars in the Galaxy, that opens up the prospect for many SNe Ia to have the same origin.

Based on observations obtained at the Southern Astrophysical Research (SOAR) telescope (NOIRLab Prop. 2022A-606104), which is a joint project of the Ministerio da Ciencia, Tecnología e Innovaciones (MCTI/LNA) do Brasil, the US National Science Foundation's NOIRLab, the University of North Carolina at Chapel Hill (UNC), and Michigan State University (MSU). We are grateful to the SOAR staff for their help in performing the observations of this project.

This paper includes data gathered with the 6.5 m Magellan Telescopes located at Las Campanas, Chile.

This work has made extensive use of the Gaia EDR3. Gaia data are being processed by the Gaia Data Processing and Analysis Consortium (DPAC). Funding for the DPAC is provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement (MLA). The Gaia mission website is https://www.cosmos.esa.int/gaia. The Gaia archive website is https://archives.esac.esa.int/gaia. This work was (partially) funded by the Spanish Ministry of Science and Innovation (MICINN), the Agencia Estatal de Investigación (AEI) 10.13039/501100011033 and by "ERDF A way of making Europe" by the European Union through grants RTI2018-095076-B-C21 and PID2021-122842OB-C21, and the Institute of Cosmos Sciences University of Barcelona (ICCUB, Unidad de Excelencia 'María de Maeztu') through grant CEX2019-000918-M.

P.R.-L. also acknowledges support from grant PGC2018-095157-B-I00, from the MICINN. J.I.G.H. acknowledges financial support from the MICINN PID2020-117493GB-I00, and also from the Spanish MICINN 2013 Ramón y Cajal program RYC-2013-14875. R.C acknowledges financial support from grant PGC2018-095157-B-I00 from the Spanish MICINN. L.G. acknowledges financial support from the Spanish MICINN, AEI10.13039/501100011033, and the European Social Fund (ESF) "Investing in your future" under the 2019 Ramón y Cajal program RYC2019-027683-I and the PID2020-115253GA-I00 HOSTFLOWS project, from Centro Superior de Investigaciones Científicas (CSIC) under the PIE project 20215AT016, and the program Unidad de Excelencia María de Maeztu CEX2020-001058-M.

P.R.L. would like to thank Evan Bauer and Warren Brown at the Harvard–Smithsonian Center for Astrophysics, and Christian Knigge from the University of Southampton, for conversations. The authors would like to thank the anonymous referee for the valuable comments on the manuscript.

The analysis of the combined high-resolution MIKE spectrum (R ∼ 28,000) of the target star MV-G272 is depicted in Figure 17, where we have zoomed into several spectral regions from the whole spectral range shown in Figure 12. We also display the residuals the observed minus computed synthetic spectra (O-C) to compare the observations and the model. We see some remaining features corresponding to residuals coming from sky subtraction and the telluric spectrum.

We have tested our stellar parameter and metallicity determination using high-resolution CARMENES-VIS spectra (Reiners et al. 2018) with a resolution of R ∼ 94,600, degraded to a resolving power of R ∼ 28,000. We chose two M1V stars Karmn J00183+440 (GX And) and Karmn J05415+534 (HD 233153), with injected white noise down to S/N ∼ 10, as explained in Section 6.

Zoom In Zoom Out Reset image size

A.1. CARMENES Spectra

We display in Figures 18 and 19 the resulting spectra of these two M dwarf stars compared with a synthetic spectrum in the same spectral range as the MIKE spectrum of the target star MV-G272 displayed in Figure 12.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size