Target EC20058-5234 Vmag 15.6 RA 20 09 40 (2000) DEC -52 25 22 (2000)

(EC20058-5234 has 2 nearby companions 2 and 4 arcsec distant with V=15.7 and V=17.8. Neither emits significant flux in the B band. The total magnitude of the 3 stars is V=14.9.)

White dwarf stars are the end points of the lives of most stars. Their study promises to provide strong constraints on the prior evolution of their progenitors. The study of the pulsating degenerates has already opened new windows of understanding in the structure of white dwarfs in general (e.g. Winget et al. 1991, 1994). One of the mysteries of the white dwarfs is the bifurcation of the majority of the population into DA (with hydrogen-dominated photospheres) and DB (with helium-dominated photospheres) white dwarfs. The target in this proposal, EC20058-5234, is a pulsating DB white dwarf.

The study of GD358 by Winget et al. (1994) has provided the most complete picture of a pulsating DB white dwarf (DBV) to date. The stellar mass, He layer mass, luminosity and distance were all derived from the l=1 triplets in the power spectrum. In addition, differential rotation and a magnetic field were also detected from their effect on the rotational splitting. The work of Watson, Klumpe and their co-workers suggests that most of the periods seen in the other known DBVs can be fit into a pattern of g-modes with l=1. The shortest period seen, ~400 s, corresponds to k=8.

Recently, Koen et al. (1995, MNRAS, in press) have discovered a new DBV, EC20058-5234, from the Edinburgh-Cape survey. This star is the second brightest DBV and, remarkably, has 7 definite periods in the range 111-351 s. The amplitude of these oscillations is very small. EC20058 is thus the first low amplitude, short period DBV, analogous to the low amplitude, short period DAVs (e.g. Clemens 1994). Koen et al. were unable to fit the observed periods into a mode identification scheme with a single value of l; it is possible, perhaps even likely, that a mixture of l's is needed. It is also likely that one of the shortest periods seen in EC20058-5234 has k=1 or 2.

An asteroseismological investigation into EC20058 thus holds the promise of different and complementary insight into the DBVs, compared to that gained from the study of GD358:

• The likely mixture of different values of l, along with the expected rotational splitting, will allow unambiguous l identification as was achieved for PG1159-035 (Winget et al. 1991). Asteroseismological mining of the star can then begin in earnest with the potential to provide fundamental parameters such as the stellar mass, luminosity, distance, rotation rate and presence or otherwise of magnetic fields (Winget et al. 1991, 1994).
• g-modes with lower values of k will sample a different, deeper region of the envelope of the star, compared to that sampled in GD358.
• Study of this star will offer insight into the surprisingly low He layer mass found for GD358 ~10^(-6) M* by providing a measurement in another star.

In order to exploit this potential, mode identification much be achieved and for this, additional modes must be identified. The aim is to ``fill in'' the frequency intervals lacking strong modes which prevented Koen et al. from identifying the value of l of the modes observed so far. Furthermore, as was shown so convincingly for PG1159-035 (Winget et al. 1991), the detection of rotational splitting provides invaluable assistance in the assignment of l. It is only when the value of l of a sequence of modes of the same k is identified that the structure of the star can be inferred from the models. Furthermore, for comparison with GD358, the longest reliably-determined period in EC20058 is 351 s, shorter than the shortest period, 423 s, of a normal mode seen in GD358. This ``gap'' must also be filled in.

To illustrate these arguments, the following table shows the observed periods in EC20058 listed by Koen et al., compared with those derived from a representative model of a DBV, kindly provided by Dr. Paul Bradley of Los Alamos National Lab.

Observed periods in EC20058:} 111, 134, 195, 204, 257, 281, 333, 351 s.

Some interesting matches can be found but the fit is far from perfect. Bearing in mind that the model periods can not only be ``scaled'' by unknown parameters such as the stellar mass, their relative spacing can also be affected by mode trapping which affects the period of each mode differently. Thus, a much richer set of observed periods is required before theoretical interpretation can be constrained. This, together with the search for rotational splitting to confirm the assignment of l to each mode, requires the high signal-to-noise and alias-free power spectra that the Whole Earth Telescope provides.

References:

Clemens J.C., 1994. PhD thesis, University of Texas at Austin.

Winget D.E., et al., 1991. ApJ, 378, 326

Winget D.E., et al., 1994. ApJ, 430, 839