HD 271791: an extreme supernova run-away B star escaping from the Galaxy1footnote 11footnote 1Based on observations obtained at the European Southern Observatory, proposal 073.D-0495(A). (2024)

1. Introduction

Studies of stellar dynamical processes in the vicinity of a super-massiveblack hole (SMBH) predicted the existence of so-called hyper-velocitystars (HVSs), objects moving faster than the Galactic escapevelocity (Hills, 1988; Yu & Tremaine, 2003). Besides being indicators for theproperties of the SMBH in the Galactic Centre (GC) and the surrounding stellar population,HVSs are also regarded as valuableprobes for the shape of the Galactic dark matter halo because of theunique kinematic properties of stars expelled from the GC (Gnedin et al., 2005; Yu & Madau, 2007).

The first HVSs have onlyrecently been discovered serendipitously(Brown et al., 2005; Edelmann et al., 2005; Hirsch et al., 2005; Heber et al., 2008, Paper I).A total of eleven faint blue stars at high Galactic latitude have been identifiedas HVSs so far after a systematic search (see Brown et al., 2007).

All except two were found to be compatible with a scenario of ejection by theSMBH in the Galactic centre (GC). One exception is the early B-typestar HE 0437$-$5439 (catalog HE 0437-5439) which probably originatesin the Large Magellanic Cloud (LMC, Przybilla et al.2008), possibly expelleddynamically by interaction of massive binaries in a dense star cluster(Leonard, 1991) or by an intermediate-mass black hole (Gualandris & Portegies Zwart, 2007),as no SMBH is known to exist in the LMC.

A measurement of the proper motion of the bright (V $=$ 12.26 mag) earlyB-type giant HD 271791 (catalog HD 271791) allowed the Galactic rest-framevelocity to be determined to lie within the range 530 – 920 km s^-1,exceeding the local Galactic escape velocity and thus qualifying thestar as a HVS – and the place of birth to be constrained to the outer rimof the Galactic disk at a Galactocentric distance of $\gtrsim$15 kpc(Paper I). This left the question of the ejection mechanism to be answeredas models invoking the SMBH in the GC are ruled out.

2. Analysis

HD 271791 was observed in early 2005 at the European Southern Observatoryon La Silla, Chile, using the FEROS spectrograph on the 2.2 m telescope(Paper I).Four individual spectra with resolving power $\lambda/\Delta\lambda$ $=$ 48 000 werecoadded and smoothed over 5 pixels for the present work, resulting in a S/N of $\sim$120 to$\sim$160 in the blue visual range.

The quantitative analysis of the spectrum of HD 271791 was carriedout following the hybrid NLTE approach discussed by Nieva & Przybilla (2007, 2008)and Przybilla et al. (2006).In brief, line-blanketed LTE model atmospheres were computed with ATLAS9and ATLAS12 (Kurucz, 1993, 1996) – the latter in order toaccount for the chemical peculiarity of the star. NLTE line-formationcalculations were performed using updated versions of Detail andSurface (Giddings, 1981; Butler & Giddings, 1985). State-of-the-art model atomswere adopted (see Table1), which allow absolute elementalabundances to be obtained with high accuracy.

Multiple independent spectroscopic indicators are considered simultaneouslyfor the determination of the atmospheric parameters, effective temperature$T_{\rm eff}$ and (logarithmic) surface gravity $\log g$: all Stark-broadened Balmer linesand simultaneously four metal ionization equilibria, of SiII/III,OI/II, SII/III and FeII/III. The resulting redundancy helpsto avoid systematic errors. The microturbulent velocity $\xi$ was determinedin the standard way by demanding abundances to be independent of lineequivalent widths.Elemental abundances and the (projected)rotational velocity $v\sin i$ were determined fromfits to individual line profiles. Fundamental stellar parameters such asmass $M$, radius $R$ and luminosity $L$, as well as the evolutionaryage $\tau_{\rm evol}$ were constrained by comparison with stellar evolutiontracks, in analogy to Paper I.

The results are summarized in Table1 and a comparison of thefinal model spectrum with many strategic regions of the observed spectrumis made in Fig.1. An excellent match between the observed andsynthetic spectrum is obtained from the large scale to minute details.Our stellar parameters for HD 271791 are consistent with earlierwork (Paper I, Kilkenny & Stone, 1988), but more precise. This applies in particularto elemental abundances which have uncertainties of only about20% (statistical 1$\sigma$ standard deviation).

The uncertainties in the atmospheric parameters were constrained from thequality of the simultaneous fits to all diagnostic indicators. Statisticaluncertainties for metal abundances were determined from theindividual line abundances.Systematic errors in the abundances dueto uncertainties in atmospheric parameters, atomic data and the quality of thespectrum are about 0.1 dex (Nieva & Przybilla, 2008; Przybilla et al., 2006), i.e. as largeas the statistical errors.

Elemental abundances of HD 271791 (averaged over all lines of an element)are displayed in Fig.2, relative to the solar standard.All abundances are within a factor of 2 relative tosolar, confirming the findings fromPaper I.However, the higher accuracyof the present work allows an underlying pattern in theelemental abundances to be uncovered, which is atypical for a youngPopulation I star. The Fe abundance is subsolar by0.27 dex, indicating that the star was formed in a metal-poor environment.This is consistent with the outer metal-poorrimofthe Galactic discsuggested by the star’s kinematic properties (Paper I).The $\alpha$-process elements O, Ne, S and in particular Si are enhanced,with the exception of Mg, which along with C and N is compatible with theiron underabundance by about a factor 2 relative to solar.This is the first detection of $\alpha$-enhancement in a massive staroutside the GC.

$\alpha$-enhancement is a characteristic of material ejected in thesupernova (SN) explosion of a massive star. In-situ formation of theseelements in HD 271791 via hydrostatic nuclear burning and subsequentmixing into the stellar atmosphere can be excluded. Typically, only low-massstars of the old Population II and the GC population show a chemicalcomposition with $\alpha$-enhancement in the Galaxy at present. Membership ofthe massive B-star HD 271791 to either population is ruled out. Note thatHD 271791 cannot be of GC origin (where[Fe/H] $\approx$ 0 and [$\alpha$/Fe] $\approx$ 0.3,e.g. Cunha et al.2007) by reason of its chemicalsignature in addition to its kinematics.

However, contamination of the surface layersof a star in a binary system is possible when the companion explodes ina SN, as reported e.g. for the secondary of the low-mass X-ray binaryNova Scorpii 1994 (catalog GRO J1655-40) (Israelian et al., 1999). A scenario is thereforedeveloped inwhich the violent disruption of a massive binary by a SN-like eventaccounts for both, the $\alpha$-enhancement of the atmosphere ofthe surviving secondary (HD 271791) and its acceleration to hyper-velocity.

3. Binary scenario for accretion of SN ejecta and acceleration

Any plausible progenitor system has to include a very massiveprimary, as the travel time of HD 271791 from the Galactic discto its present location in the halo requires the SN-like event tohappen early in the star’s lifetime. Stars above 40 $M_{\odot}$die within less than 6 Myrs (Meynet & Maeder, 2005), which is sufficiently shortfor this scenario.

Mass accretion from a SN-shell by the secondary dependsstrongly on geometric factors because of the high velocity of the ejectacompared to the escape velocity from thesurface of the secondary (several 10³ vs. $\sim$10³ km s^-1). Interaction of the SN-shellwith the secondary is therefore possible for a fraction of mass equal to thesolid angle subtended from the primary at most (Fryxell & Arnett, 1981, FA81), unless thereis significant fall-back of material onto the remnant.A close binary system is therefore favoured, similar to the scenario developedfor Nova Sco 1994 (Podsiadlowski et al., 2002, see in particular their Fig.2).In the present case, both components mustbe massive and the explosion of the primary has to be strongly asymmetric todisrupt the system and to release the secondary at its orbital velocity.

The absence of N enrichment indicates that HD 271791 did not accretesignificant amounts of (CN-processed) mass from the primary, i.e. the systemavoided Roche-lobe overflow (Vanbeveren et al., 1998, V98). This implies that the systemhas undergone a common-envelope phase, with spiral-in of the secondary becauseof viscous forces. Systems with mass ratios (secondary/primary mass)$q$ $\leq$ 0.2 qualify for a wide range of initial periods(V98), resulting in a minimum mass of $\sim$55 $M_{\odot}$ forthe primary on the zero-age main sequence.

The spiral-in process can expel the H-rich envelope of the primary(stopping the merger) when sufficient orbital energy is deposited, at the expenseof a very large reduction of the orbital separation (V98).Accordingly, a binary composed of a Wolf-Rayet (WR) $+$ an early B-type main-sequencestar is a conceivable progenitor system for the SN-like event involvingHD 271791. The radius of an 11 $M_{\odot}$ star is $\sim$4 $R_{\odot}$on the main sequence, and $\sim$2-3 $R_{\odot}$ for a WR star with mass$\leq$ 20 $M_{\odot}$ of WC subclass at the baseline metallicity ofHD 271791 (Crowther et al., 2002). As a consequence, the separation $a$ between the binarycomponents before the SN-like event could be as small as 13-15 $R_{\odot}$ for adetached system, resulting in an orbital period of $\sim$1 day and an orbital velocityof the secondary of $\sim$400 km s^-1.

At this point a first phase of accretion onto HD 271791 is likely to occur.The WR star has a dense and strong wind that will directly impact on thesurface of the secondary, as a metal-poor B1 V star has no significant wind of its own(a similar, but more extreme situation to that observed for the massive binaryCPD $-$41° 7742, Sana et al., 2005). Using conservative assumptions on WR mass-loss rates and lifetime,we estimate that about 0.04 $M_{\odot}$ of material,rich in C and to a lesser degree abundant in O and Ne, could be deposited.

The final collapse of the WC star into a black hole (Heger et al., 2003) results in anordinary type Ic SN ora more energetic hypernova (HN). The kick experienced by the (proto-) black hole disruptsthe binary and releases the companion at its orbital velocity, in analogy to the classicalscenario of run-away stars (Blaauw, 1961).

This gives rise to the second – and main – mass accretion event onto HD 271791.The interaction of the expanding SN-shell with the secondary is acomplex hydrodynamical process involving ablation of mass from the outerlayers of the secondary and accretion from the shell, which is not completelyunderstood (FA81). Also interaction with fall-back material from theexplosion can be significant (Podsiadlowski et al., 2002). We therefore made conservative assumptionsto estimate the effects of accretion of SN-processed material on the chemicalcomposition of the secondary. Nucleosynthesis yields of Nomoto et al. (2006) were used.We chose the highest mass models available, for a star of initially40 $M_{\odot}$ and metallicity $Z$ $=$ 0.004,that comes closest to the adopted pristine value. The total kinetic energyof the explosion predicted by the models is 3$\times$10⁵² erg for the HN and10⁵¹ erg for the SN, respectively. A total of $M_{\rm acc}$ $=$ $M_{\rm ejecta}\,R_{2}^{2}/4a^{2}$may be accreted by the secondary with radius $R_{2}$.However, due to the large momentum of the impacting material the efficiency of accretionis small, i.e. $\sim$10% (FA81).About 0.02 $M_{\odot}$ of heavy elements can therefore be accretedin the event that expels $M_{\rm ejecta}$ $\approx$ 10 $M_{\odot}$ ofmetals in total (for $a$ $=$ 13 to 15 $R_{\odot}$).At the same time $\sim$1% (0.1 $M_{\odot}$) of the secondarymass is ablated from the surface layers (FA81), including a considerable partof the material accreted previously from the WR wind.

The accreted material has been mixed with unpolluted matter fromthe secondary’s deeper layers and possibly with CN-processed from its coreover the past $\sim$20 Myrs. Chemical mixing in the radiative envelopesof rotating stars is slow, approximated well by diffusion(Maeder & Meynet, 2000). The metal-rich material will therefore be only partially mixed,leading to a dilution of the original HN/SN abundance pattern.

In order to model the observed abundance pattern, several parametershave to be adjusted: i) the initial chemical composition; ii) the fractionof the SN-shell that is accreted and iii) the equivalent envelope massaffected by complete mixing (in reality the heavy elementabundances will be stratified). The scenario has to reproduce the observed Fe abundance as aboundary condition as the abundance patterns are normalised with respect to iron.The model abundance patterns do not account for the survivingfraction of the C-rich and N-free WC wind and rotational mixing with slightly C-depletedand N-enriched material from the core as quantitative estimates are difficult to make.However, both processes tend to cancel each other, so that the net effect isexpected to be small.

Results are shown in Fig.3, for themixing of 0.12%/0.08% of the HN/SN ejecta with 1 $M_{\odot}$ envelopematerial of HD 271791of pristine composition [$X$/H] $=$ $-$0.4/$-$0.24 dex.Note that mixing of a higher fraction of envelope mass can compensate for an increase of theaccreted mass and vice versa.Overall, reasonable agreement between the model predictions and observation isobtained, except that too large a Mg and too low a Siabundance is predicted. However, this is likely an artifact of the use ofintegrated yields. More realistic is a latitudinal dependency of the ejectacomposition in the likely case of an aspherical explosion (Maeda et al., 2008)and a stratified abundance distribution in the ejecta.Hydrodynamic instabilities at the trailing edge of theSN-shell will preferentially lead to an accretion of material from theinner shell on the secondary (FA81), which is rich in Si and depleted in Mg(e.g. Maeda et al., 2002). Tailored nucleosynthesis calculations for thisparticular case, a detailed hydrodynamical simulation of theSN-shell–secondary interaction and a comprehensive modeling of thesubsequent mixing of the accreted material with the secondary’s envelope aredesirable.

4. Conclusions

An accurate determination of the surface composition of the HVS HD 271791in the present work confirmed the low [Fe/H] expected for a star born in the outerGalactic rim (Paper I). The finding of $\alpha$-enhancement allowed theacceleration mechanism to be identified as an extreme case of the run-awayscenario (Blaauw, 1961).The evolution of a binary with a verymassive primary ($M$ $\gtrsim$55 $M_{\odot}$) can lead to a suitableprogenitor system of a close WR $+$ B main sequence star that can accountfor both, the observed $\alpha$-enhancement and ejection at high orbitalvelocity. The collapse of the WR star into a black hole gave rise to eitheran (aspherical) SN or a more energetic HN (which may be accompanied bya gamma-ray burstof the long-duration/soft-spectrum type, Woosley & Bloom, 2006), in which the systemwas disrupted. A simple model for accretion from the SN-shell and subsequent mixing withpristine envelope material can qualitatively explain the observedabundance pattern of HD 271791. Only one fact remains to be explained:the difference between the kick velocity from the binary disruption($v_{\rm kick}$ $\approx$ 400 km s^-1) and the Galactic rest frame velocity of HD 271791($v_{\rm grf}$ $=$ 530–920 km s^-1).

Inspection of Fig.4 in Paper I shows that HD 271791 was ejected from theMilky Way in the direction of the Galactic rotation. A rough alignment ofthe orbital velocity vector of the secondary with the Galactic rotation directionat the moment of system disruption is therefore sufficient to explain thespace motion of HD 271791 if it is on the lower end of the observed range.Numerical kinematic experiments using theGalactic potential of Allen & Santillan (1991) imply a minimum $v_{\rm kick}$ $\approx$ 380 km s^-1 to obtain the minimum $v_{\rm grf}$of HD 271791 ($\sim$530 km s^-1) that is consistent with the observedproper motion, radial velocity, distance and age. For the ‘best’ proper motion of Paper I($v_{\rm grf}$ $=$ 630 km s^-1) a $v_{\rm kick}$ $\approx$ 460 km s^-1 would be required.We conclude that the proposed SNrun-away scenario can reproduce all observational constraints, thekinematics and the abundance pattern. Extreme cases of the run-away scenarioof Blaauw (1961) therefore pose an alternative to the Hills mechanism forthe acceleration of some HVSs with moderate velocities.As HD 271791 is escaping from the Galaxy, it could be termed ahyper-runaway star, the first of its class.

We thank R. Napiwotzki for valuable discussions on the kinematics ofHD 271791 and the ESO staff at La Silla for support with the observations.M.F.N. gratefully acknowledges financial support by the DeutscheForschungsgemeinschaft (grant HE 1356/45-1).

References

• Allen & Santillan (1991)Allen, C., & Santillan, A. 1991, Rev. Mexicana Astron. Astrofis., 22, 2552
• Becker (1998)Becker, S. R.1998, ASP Conf. Ser., 131, 137
• Becker & Butler (1988)Becker, S. R., & Butler, K. 1988, A&A, 201, 232
• Becker & Butler (1990)Becker, S. R., & Butler, K. 1990, A&A, 235, 326
• Blaauw (1961)Blaauw, A.1961, Bull. Astr. Inst. Neth., 15, 265
• Brown et al. (2005)Brown, W. R., Geller, M. J., Kenyon, S. J., & Kurtz, M. J.2005, ApJ, 622, L33
• Brown et al. (2007)Brown, W. R., Geller, M. J., Kenyon, S. J., et al. 2007, ApJ, 671, 1708
• Butler & Giddings (1985)Butler, K. & Giddings, J.R.1985, in Newsletter of Analysis of Astronomical Spectra, No. 9 (Univ. London)
• Crowther et al. (2002)Crowther, P. A., Dessart, L., Hillier, D. J., et al. 2002, A&A, 392, 653
• Cunha et al. (2007)Cunha, K., Sellgren, K., Smith, V. V., et al. 2007, ApJ, 669, 1011
• Dufton et al. (1986)Dufton, P. L., et al. 1986, MNRAS, 222, 713
• Edelmann et al. (2005)Edelmann, H., Napiwotzki, R., Heber, U., et al. 2005, ApJ, 634, L181
• Fryxell & Arnett (1981)Fryxell, B. A., & Arnett, W. D.1981, ApJ, 243, 994 (FA81)
• Giddings (1981)Giddings, J. R.1981, Ph.D. Thesis, University of London
• Gnedin et al. (2005)Gnedin, O. Y., et al. 2005, ApJ, 634, 344
• Grevesse & Sauval (1998)Grevesse, N., & Sauval, A. J.1998, SpaceSci.Rev., 85, 161
• Gualandris & Portegies Zwart (2007)Gualandris, A., & Portegies Zwart, S. 2007, MNRAS, 376, L29
• Heber et al. (2008)Heber, U., et al. 2008, A&A, 483, L21 (Paper I)
• Heger et al. (2003)Heger, A., Fryer, C. L., Woosley, S. E., et al. 2003, ApJ, 591, 288
• Hills (1988)Hills, J. G.1988, Nature, 331, 687
• Hirsch et al. (2005)Hirsch, H. A., Heber, U., O’Toole, S. J., & Bresolin, F.2005, A&A, 444, L61
• Israelian et al. (1999)Israelian, G., Rebolo, R., Basri, G., et al. 1999, Nature, 401, 142
• Kilkenny & Stone (1988)Kilkenny, D., & Stone, L. E.1988, MNRAS, 234, 1011
• Kurucz (1993)Kurucz, R. L.1993, CD-ROM 13 (Cambridge: SAO)
• Kurucz (1996)Kurucz, R. L. 1996, ASP Conf. Ser., 108, 160
• Leonard (1991)Leonard, P. J. T. 1991, AJ, 101, 562
• Maeda et al. (2002)Maeda, K., Nakamura, T., Nomoto, K., et al.2002, ApJ, 565, 405
• Maeda et al. (2008)Maeda, K., Kawabata, K., Mazzali, P. A., et al. 2008, Sci, 319, 1220
• Maeder & Meynet (2000)Maeder, A., & Meynet, G. 2000, ARA&A, 38, 143
• Meynet & Maeder (2005)Meynet, G., & Maeder, A.2005, A&A, 429, 581
• Morel & Butler (2008)Morel, T., & Butler, K. 2008, A&A, in press, arXiv:0806.0491 [astro-ph]
• Morel et al. (2007)Morel, T., Butler, K., Aerts, C., et al. 2007, A&A, 457, 651
• Nieva & Przybilla (2006)Nieva, M. F., & Przybilla, N. 2006, ApJ, 639, L39
• Nieva & Przybilla (2007)Nieva, M. F., & Przybilla, N.2007, A&A, 467, 295
• Nieva & Przybilla (2008)Nieva, M. F., & Przybilla, N.2008, A&A, 481, 199
• Nomoto et al. (2006)Nomoto, K., Tominaga, N., Umeda, H., et al. 2006, Nucl.Phys.A, 777, 424
• Podsiadlowski et al. (2002)Podsiadlowski, P., Nomoto, K., Maeda, K., et al. 2002, ApJ, 567, 491
• Przybilla (2005)Przybilla, N. 2005, A&A, 443, 293
• Przybilla & Butler (2001)Przybilla, N., & Butler, K.2001, A&A, 379, 955
• Przybilla & Butler (2004)Przybilla, N., & Butler, K.2004, ApJ, 609, 1181
• Przybilla et al. (2000)Przybilla, N., Butler, K., Becker, S. R., et al. 2000, A&A, 359, 1085
• Przybilla et al. (2001)Przybilla, N., Butler, K., Becker, S. R., Kudritzki, R. P. 2001, A&A, 369, 1009
• Przybilla et al. (2006)Przybilla, N., Butler, K., Becker, S. R., Kudritzki, R. P. 2006, A&A, 445, 1099
• Przybilla et al. (2008)Przybilla, N., Nieva, M. F., Heber, U., et al.2008, A&A, 480, L37
• Sana et al. (2005)Sana, H., Antokhina, E., Royer, P., et al. 2005, A&A, 441, 213
• Vanbeveren et al. (1998)Vanbeveren, D., et al. 1998, A&ARev., 9, 63 (V98)
• Vrancken et al. (1996)Vrancken, M., Butler, K., & Becker, S. R.1996, A&A, 311, 66
• Woosley & Bloom (2006)Woosley, S. E., & Bloom, J. S.2006, ARA&A, 44, 507
• Yu & Madau (2007)Yu, Q., & Madau, P.2007, MNRAS, 379, 1293
• Yu & Tremaine (2003)Yu, Q., & Tremaine, S.2003, ApJ, 599, 1129