Sample light curve of PG~1336 showing primary and secondary eclipses.
Subdwarf B (sdB) stars are extreme horizontal branch stars which are believed to be comprised of 0.5 solar masses of He (and the usual admixture of metals) with a very thin (less than 2\% by mass) layer of H, chemically segregated from the He and heavier elements by settling in the star's large gravitational field. The evolution of the progenitors of these stars is not well understood, relying on unknown mass loss mechanisms on the giant branch to remove almost all the H envelope of the progenitor at the same time as increasing the He core to the mass needed to ignite He burning in the He flash. Fortunately the discovery in the past few years of pulsating sdB stars (the EC14026 stars) (Kilkenny et al. 1997) has opened the prospect of using asteroseismology to examine the interiors of these stars. Understanding this stage in stellar evolution is important for distance calibrations (being similar to RR~Lyrae stars) and population synthesis studies (to explain the UV excesses in globular clusters and elliptical galaxies).
PG1336 is a partially eclipsing binary that pulsates in at least 15 modes (Reed et al 1999). It has a binary period of ~2.4 hours and the companion is an M4-5 star of approximately the same radius but considerably fainter (Kilkenny et al. 1998). As such, the companion contributes little light to the system (except via a strong reflection effect)(see Figure 1), yet the orbit is close enough to assume that it is tidally locked. This provides us with a known rotation period, from which we can search for rotationally split modes, a sign of non-radial pulsations. At an inclination of ~81 degrees, we effectively view the stars equator-on and the primary eclipse covers approximately half of the sdB star. Because of this changing viewing geometry the pulsation amplitudes will also change through the eclipse. In an equator-on system, l=1, m=0 modes should show reduced net amplitude because of geometric cancelation across the star's surface (see Figure 2). However, during eclipse, this mode is rendered visible as one hemisphere is temporarily blocked. The same is true for l=2, m=+/-1 modes. While the amplitude of nearly every non-radial mode will be affected (some minimally), any radial modes will remain unaffected. As such, PG1336 offers a unique opportunity to identify pulsation modes in an sdBV star.
We propose to observe PG1336-018 with the Whole Earth Telescope (WET). The WET (Nather et al. 1990) is a collaboration of astronomers who agree to observe a target star in such a way that observations can continue around the globe, from site to site, 24 hours a day. The continuous coverage offered by the WET is the only way to separate real rotationally-split modes from aliases caused by data gaps, or in this case, orbital sidebands. The increased data density also provides a significantly higher signal to noise ratio than is possible from a single site observing over a similar timespan. We find we typically need 2 weeks of coverage from the network to obtain a resolution and signal to noise ratio high enough for accurate mode identification. With a Vmag of 13.4, PG1336 is easily visible from 1m class telescopes, but the real observed signal is not the mean level of the star, but the deviations around that mean due to the stellar pulsations. With the fairly low amplitudes of this star (we are expecting modes of 0.003 mag or less), 2m class telescopes are really more appropriate. This class of telescope also allows us a higher time resolution through the 900 second primary eclipse. We are requesting time at several sites around the globe, including: Beijing Astrophysical Observatory 2.16m, SAAO 1.9m, McDonald 2.1m, Observatoire de Haute Provence 1.9m, SAAO 1.9m, Moletai 1.65m, Itajuba 1.6m, CTIO 1.5m, Calar Alto 1.2m, Mt. John 1.0m, Siding Spring 1.0m, Naini Tal 1.0m, Wise Observatory 1.0m, and Mt. Suhora 1.0m.
Our goals for this WET run are to: 1) Obtain 2 weeks of continuous coverage of PG1336 to beat down daily aliases inherent in single-site data. 2) Use the known orbital period, assume the stars are tidally locked and search for rotationally split modes as a sign of non-radial pulsations. 3) With an inclination of 81 degrees, the primary eclipse covers roughly half of the pulsating star. This allows us to observe modes (during eclipse) that we would not see otherwise. 4) Compare the amplitudes of pulsation seen in-eclipse to those out of eclipse to make a positive identification of pulsation modes, which may then be used to constrain models.
It is very likely that, in a binary this close, the rotation of the pulsator will be phase-locked to the orbital motion. If so, non-radial pulsations will be split according to the Ledoux formula: $ f_{klm} = f_0 + m*(1-C_{lk})*F_{rot}$ with Frot=1/2.4hr, the star's rotation frequency. The rotational splitting coefficient C(kl) is a function of the structure of the star, and is estimated from pulsation models for the sdB stars developed by Kawaler (ISU) to be between 0.05 and 0.4, significantly NON-zero. It is inevitable from the shape of the light curve (eclipses, reflection effect) that a frequency splitting equal to the orbital frequency will be seen in the pulsations. But rotational splitting would be resolved from this in ~3-20 binary periods provided C(kl) is in the expected range. Since these stars pulsate in a range where we expect to see both radial and non-radial pulsations, detection of rotational splitting would aleviate some of this ambiguity.
The data from such a WET run will have some additional complexities over normal observations of variable stars. However, we already have software in place that has been tested. We are thus able to effectively remove the reflection effect from the light curve as well as extract and flatten the primary eclipse for use in pulsation analysis (Figure 3). Where other multi-site campaigns have found an abundance of pulsation modes in pulsating sdB stars, this WET campaign may be the first real opportunity to identify modes, rather than just detect them.
Kawaler, Steven D., 1999, A.S.P. Conf. Ser. Vol 169, 158
Kilkenny D., et al., 1997, MNRAS, 285, 640
Kilkenny D., et al., 1998, MNRAS, 296, 329
Nather, R.E., et al., 1990, ApJ, 361, 209
Reed, M.D. et al., 1999, Balt.A. 9, 183
Winget D.E., et al., 1991, ApJ, 378, 326
Winget D.E., et al., 1994, ApJ, 430, 839
Sample light curve of PG~1336 showing primary and secondary eclipses.
Sample model images (at 200 second intervals) and light curves for
simulations of PG~1336.
Right: Light curves from simulations of non-radial pulsations. The pulsations
have been corrected by the constant curve on the bottom.
Top: Extracted eclipse with correcting contour (line) offset by -0.2.
Bottom: Flattened eclipse showing obvious pulsations.
Long justification: latex format.
Long justification: ps format.
Short justification: html format.
Short justification: ps format.
Short justification: latex format.
Figure 1: ps format.
Figure 2: ps format.
Figure 3: ps format.
Figure 4: ps format.