1RXS J0823.6-2525: a new ultramassive magnetic white dwarf

Mon. Not. R. Astron. Soc. 299, L1–L4 (1998)

1RXS J0823.6–2525: a new ultramassive magnetic white dwarf Lilia Ferrario, Ste´phane Vennes and Dayal T. Wickramasinghe Department of Mathematics and Astrophysical Theory Centre, Australian National University, Canberra, ACT 0200, Australia

Accepted 1998 June 17. Received 1998 May 6

A B S T R AC T

We present observations and an analysis of the X-ray source 1RXS J0832.6–2525 which shows it to be a low field magnetic white dwarf with an unusual high mass. This is the second magnetic white dwarf for which a determination of a spectroscopic mass has been possible, and both stars belong to the growing class of ultramassive white dwarfs (M $ 1:1 M( ). Key words: stars: individual: 1RXS J0823.6–2525 – stars: magnetic fields – white dwarfs.

1

INTRODUCTION

Mass determinations of magnetic white dwarfs have proven to be difficult because of the complex manner in which Balmer lines are split by the Zeeman effect at high fields, and by the dominance of magnetic field broadening over pressure broadening. The former depends on magnetic field spread which, in turn, depends on field geometry and viewing aspect, while the latter depends on gravity and temperature. As a result, the best available mass estimates are based on effective temperatures and distances based on parallax measurements. Liebert (1988) first noted that magnetic white dwarfs tended to have a lower absolute luminosity than their nonmagnetic counterparts with similar colours. Direct spectroscopic mass determinations of magnetic white dwarfs may be attempted, particularly in low-field objects; Schmidt et al. (1992) measured the surface gravity of the 3.5-MG DA white dwarf PG 1658+441 (log g = 9.36) indicative of a mass of 1.31 M( . The extreme ultraviolet (EUV)-selected sample of white dwarfs yielded some 10 ultramassive white dwarfs ($1:1 M( ) out of ,110 objects (Vennes et al. 1997), and, interestingly, the only two magnetic white dwarfs in the sample appear to be ultramassive [EUVE J0317–855 (Ferrario et al. 1998; Barstow et al. 1995) and PG 1658+441]. Few such objects are known, and population characteristics are uncertain. A search for new massive and magnetic white dwarfs has recently been extended to hot white dwarfs detected in the ROSAT All-Sky Survey (Voges et al. 1996). We report the discovery of a new magnetic white dwarf, 1RXS J0823.6–2525, which also appears ultramassive. Section 2 presents our spectroscopy and a comparison with the DAp PG 1658+441, and the following two sections describe the stellar parameters of the new white dwarf, temperature and gravity (Section 3), and magnetic field (Section 4). We summarize in Section 5. 2

O B S E RVAT I O N S

Two spectra of the white dwarf 1RXS J0823.6–2525 were obtained with the 74-inch telescope at Mount Stromlo Observatory on 1997 December 2 and 3 as part of our optical identification campaign of southern ROSAT sources (Vennes, in preparation). The candidate star is the closest to the X-ray source position (see finder chart, q 1998 RAS

Fig. 1), and is also the candidate for EUV detections (Pye et al. 1995; Bowyer et al. 1996). We used the Cassegrain spectrograph ˚ and a 300 line mm¹1 grating providing a dispersion of 2.75 A ˚ ). We then obtained higher quality pixel¹1 (or a resolution of ,6 A spectra on 1998 January 22 with the 2.3-m telescope at Siding Spring Observatory. We used the double spectrograph (DBS) equipped with SITe 1752×532 CCDs, a 300 line mm¹1 grating on ˚ pixel¹1 ), and a 316 line mm¹1 grating on the the blue side (2.18 A ˚ red side (2.07 A pixel¹1 ). Fig. 2 shows a comparison between a spectrum of the new white dwarf and spectra of the DA GD 50 and the magnetic DAp PG 1658+441 obtained by Vennes et al. (1997). Both comparison stars are ultramassive white dwarfs, but Schmidt & Smith (1995) determined an upper limit of ,40 kG to the magnetic field of GD 50, while Schmidt et al. (1992) measured a polar field of 3.5 MG in PG 1658+441. The star 1RXS J0823.6– 2525 bears some resemblance to the latter, and the new white dwarf is clearly magnetic as well as massive. 3 E F F E C T I V E T E M P E R AT U R E A N D S U R FAC E G R AV I T Y O F 1 R X S J 0 8 2 3 . 6 – 2 5 2 5 We first measure the effective temperature and surface gravity by modelling the Balmer line series, neglecting the line cores which are affected by Zeeman splitting. We use the grid of model atmospheres for non-magnetic white dwarfs described by Vennes et al. (1997). We measured the effective temperature and surface gravity, Teff ¼ 43 200 6 1000 K and log g ¼ 9:02 6 0:10, which establishes its high mass. A comparison with Wood’s (1995) evolutionary sequences shows that 1RXS J0823.6–2525 is a 1:2 6 0:04 M( white dwarf, with a cooling age of 40 × 106 yr. The star 1RXS J0823.6–2525 is in the Galactic plane [ðl; bÞ ¼ ð246:1; þ6:79Þ] and, adopting an absolute magnitude of MV ¼ 11:31 and measuring the apparent magnitude from the optical spectrum (mV ¼ 16:4 6 0:3), we estimate its distance to 91–120 pc. Schmidt et al. (1992) estimate the mass of PG 1658+441 to 1.31 M( , and both objects belong to the class of ultramassive DA white dwarfs. 4

ZEEMAN MODELLING

The broad and structured Balmer lines seen in the spectrum of

L2

L. Ferrario, S. Vennes and D. T. Wickramasinghe

←

1RXS J0823.6-2525 Figure 1. Finder chart (7 × 7 arcmin2 ) centred on the ROSAT source position (a2000 = 08h 23m 23:s5, d2000 ¼ ¹258 250 2200 ). The new white dwarf is marked with an arrow, and a position error circle (30-arcsec radius) is drawn. North is up and east is left.

1RXS J0823.6–2525 can be explained as being a result of Zeeman structure caused by a weak surface-averaged field of about 2 MG. In this section, we will endeavour to model the observed spectral characteristics by fitting the positions and shapes of the Zeeman components to magnetic white dwarf model atmosphere calculations for an assumed field structure. At sufficiently high magnetic fields, hydrogen lines in white dwarfs are broadened not only by Stark and Doppler effects, but also by the magnetic field spread across the visible stellar disc. The modelling of spectral lines requires a careful treatment of the polarized radiative transfer equations, allowing also for magnetooptical effects, which take into account the different refractive indices for radiation with different polarization states, and which play a role in determining the depths of the lines (Martin & Wickramasinghe 1981). For our modelling, we have used the magnetic white dwarf atmosphere program developed by Wickramasinghe & Martin (1979) modified to include magneto-optical effects. The shifts and strengths in hydrogen lines, arising from the presence of magnetic fields, were included, using the results of Zeeman calculations by Kemic (1974). At low fields, the p components of the observed lines do not experience a shift in the linear Zeeman effect and they only have a relatively weak quadratic shift. In the field regime under consideration, we expect that the perturbation of the static Hamiltonian is dominated by magnetic fields with the Stark effect not affecting the positions of the Zeeman components, but broadening the Zeeman patterns. In particular, at the low fields appropriate to the present

Figure 2. Blue spectra (SSO 2.3-m, DBS) of the new DAp 1RXS J0823.6–2525 compared to spectra of the massive DAp PG 1658+441 (1.31 M( ) and DA GD 50 (1.27 M( ). In inset, an Ha spectrum (DBS) of the new DAp on the same flux scale. q 1998 RAS, MNRAS 299, L1–L4

1RXS J0823.6–2525

L3

Figure 3. Modelling of the Zeeman splitting in Balmer line series of PG 1658+441 and 1RXS J0823.6–2525. The underlying model for PG 1658+441 assumes Teff ¼ 30 000 K and log g = 9.36, and the model for 1RXS J0823.6–2525 assumes Teff ¼ 43 200 K and log g = 9.02.

star, the line wings would still be dominated by static electric fields, rather than by magnetic broadening. Accordingly, we have calculated the line wings by using zero magnetic field model atmospheres which can also be used to infer gravity, as was done in Section 3. However, the shapes of the cores of the lines are determined by a combination of Doppler broadening and magnetic field broadening caused by magnetic field spread across the stellar surface. The detailed problem of Stark broadening of hydrogen lines in a magnetic field is complex and is essentially unsolved. For example, the Stark effect is expected to depend on electric field orientation relative to the magnetic field, and can only be discerned by detailed calculations. At sufficiently low fields, in the linear Zeeman regime, it may be reasonable to assume as a first approximation that the individual Zeeman components are Stark broadened, as in the zero field case which is the approach we adopt. Our calculations have been carried out using a convolved Stark-and-Doppler profile, it being implicitly assumed that the magnetic field has no effect on the Stark broadening. The latter is calculated using the Griem (1960) theory for hydrogen line broadening. It has been shown previously (Achilleos & Wickramasinghe 1989) that it is possible to obtain useful information on magnetic field structure by a careful analysis of spectroscopic data even in the absence of polarimetry, particularly if the field is strong enough for magnetic field broadening to be important. In the present low-field case, only the cores of the lines are sensitive to field structure and field geometry, so that it is more difficult to provide strong q 1998 RAS, MNRAS 299, L1–L4

constraints on field structure. We have accordingly assumed that the field structure is that of a centred dipole. The observed splitting between the p and j components of Hb requires a mean surface field strength of about 2 MG, which can be achieved for a range of values of the dipole field strengths Bd and of the viewing angle i between the line of sight and the dipole axis. For nearly pole-on viewing, the effective field spread is larger, whereas for near-equator viewing the field is more uniform. These effects can be seen as magnetic field broadening of the line cores, and can be used to provide some constraints on the dipolar magnetic field strength Bd and the inclination angle i. The two models which bracket the observations have Bd ¼ 2:8 MG and i ¼ 208 and Bd ¼ 3:5 MG and i ¼ 608. These models are shown in Fig. 3. The zero magnetic field atmosphere was calculated using Kurucz’s (1971) ATLAS program with an effective temperature Teff ¼ 43 200 K and log g ¼ 9:02 as derived in Section 3. Fig. 3 also shows an observation of the ultramassive low-field magnetic white dwarf PG 1658+441, which exhibits a spectrum remarkably similar to that of 1RXS J0823.6–2525, and a comparison with a synthetic spectrum calculated using the model parameters derived by Schmidt et al. (1992), namely, a dipolar magnetic field strength of 3.5 MG and a viewing angle of 608 to the magnetic axis, and Teff ¼ 30 000 K and log g ¼ 9:36. Our calculations show that Hb is dominated by the linear Zeeman effect, being split into a classical Zeeman triplet as observed. Hg is also resolved into a triplet, but the splitting between the p and the

L4

L. Ferrario, S. Vennes and D. T. Wickramasinghe Table 1. Properties of the DAp 1RXS J0823.6–2525. Teff log g M B; i mV MV d

43200 6 1000 K 9:02 6 0:10 1:20 6 0:04M( 2.8–3.5 MG, 20–608 ,16:4 ,11:3 105 pc

two j components is not uniform, indicating a marginal quadratic Zeeman effect. It is to be noted that the observations do not show a resolved p component, but instead show a line with a flat-bottomed base, extending in wavelength between the calculated positions of the jþ and the j¹ components. An identical effect is seen for this line in PG 1658+441, suggesting perhaps an inadequacy in our treatment of Stark broadening of Hg . A stronger quadratic effect is seen at Hd, and a clear triplet structure is no longer evident. Instead, the line develops an asymmetric structure with the centroid of the blend moving to shorter wavelengths reflecting the blueward shift of the p and jþ components in this field regime. The wavelength of the centroid of this blend coincides with the observed central wavelength of Hd . In conclusion, we are unable to present strong constraints on field structure and orientation using magnetic white dwarf atmosphere models, although a centred dipole model appears to be adequate to represent the present data. 5

CONCLUSIONS

Table 1 summarizes the properties of the new ultramassive magnetic white dwarf 1RXS J0823.6–2525. Liebert (1988) first noted that there appeared to be a tendency for the masses of white dwarfs with strong magnetic fields to be higher than those of non-magnetic white dwarfs. The mass estimates were based on parallax measurements and included the star Grw þ7388247. Subsequently it was realized that the mass distribution of isolated white dwarfs was bimodal, with a high-mass tail peaking at 1:2 M( (Vennes et al. 1997). The new white dwarf 1RXS J0823.6–2525 belongs to the class of magnetic white dwarfs and is also ultramassive (M ¼ 1:2 M( ). Ultramassive white dwarfs tend to be peculiar. Vennes, Bowyer & Dupuis (1996) measured a helium abundance ratio He=H ¼ 2 × 10¹4 in the DA GD 50 and, based on the EUV/soft X-ray sample of white dwarfs, we now know that some 25 per cent

of massive white dwarfs are also magnetic. Schmidt & Smith (1995) found the proportion of magnetic objects among a magnitudelimited sample of white dwarfs to be around 4 per cent. These features of ultramassive white dwarfs should guide our efforts in the understanding of evolution of early main-sequence stars, likely progenitors of massive white dwarfs, or double degenerate stars which may also form massive white dwarfs after coalescence. The high mass of some magnetic white dwarfs also suggest an evolutionary link with pulsars, as most recently discussed by Schmidt & Smith (1995), and some ultramassive white dwarfs may be viewed as ‘missed’ neutron stars. AC K N O W L E D G M E N T S This work was supported by a Long Term Space Astrophysics grant to the University of California, and by an ARC QE II fellowship (SV) to the Australian National University. We gratefully acknowledge the assistance of M. Buxton and D. Christian with some of the observations. REFERENCES Achilleos N., Wickramasinghe D. T., 1989, ApJ, 346, 444 Barstow M. A., Jordan S., O’Donoghue D., Burleigh M. R., Napiwotzki R., Harrop-Allin M. K., 1995, MNRAS, 277, 971 Bowyer S., Lampton M., Lewis J., Wu X., Jelisnky P., Malina R. F., 1996, ApJS, 102, 129 Ferrario L., Vennes S., Wickramasinghe D. T., Bailey J. A., Christian D. J., 1998, MNRAS, 292, 205 Griem H. R., 1960, ApJ, 132, 883 Kemic S. B., 1974, JILA REP. No. 113 Kurucz R., 1971, Smith. Astr. Obs. Spec. Rep. No. 309 Liebert J., 1988, PASP, 100, 1302 Martin B., Wickramasinghe D. T., 1981, MNRAS, 196, 23 Pye J. P. et al., 1995, MNRAS, 274, 1165 Schmidt G. D., Bergeron P., Liebert J., Saffer R., 1992, ApJ, 394, 603 Schmidt G. D., Smith P. S., 1995, ApJ, 448, 305 Vennes S., Bowyer S., Dupuis J., 1996, ApJ, 461, L103 Vennes S., Thejll P., Genova-Galvan R., Dupuis J., 1997, ApJ, 480, 714 Voges W. et al., 1996, A&A, in press (http://www.rosat.mpe-garching. mpg.de/survey/) Wickramasinghe D. T., Martin B., 1979, MNRAS, 188, 165 Wood M. A., 1995, in Koester, D., Werner, K., eds, White Dwarfs. Springer, Berlin, p. 41 This paper has been typeset from a TE X=LA TE X file prepared by the author.

q 1998 RAS, MNRAS 299, L1–L4