Eclipsing Binary Update Number 10

Number 10 August 2000

Notes from the Chairman

In the dead of winter a preponderance of observers complain of the cloudy skies and cold temperatures and declare that their observing will pick up when summer arrives. So, then, why do fewer minima timings pass across my desk during May, June, and July? Could it be that shorter nights compounded by Daylight Savings Time (more properly known as Dark Hours Wasting Time) might have some effect? Hopefully, our southern colleagues are taking advantage of their longer nights to fill the gap.

The AAVSO Spring Meeting in Huntsville, Alabama, took on an international flavor with attendees from Australia, Canada, Switzerland, Italy, Argentina, England, South Africa, Japan, and Chile. I was especially pleased to meet with Jan Hers (South Africa) and Roger Diethelm (Switzerland), whom I had not seen for some time.

In our committee report we noted that, since August 1999, 23 observers have obtained more than ten thousand visual observations of eclipsing binaries. Five observers, led by Gerry Samolyk, made a few thousand CCD observations not yet counted. All together, data were obtained on 233 stars with nearly 470 times of minima being determined.

We also noted that preparation of the sixth monograph, listing observed times of minima, is progressing. When it is completed, it is likely that only four program stars will remain that have not yet appeared in this series of publications. These four stars, each with its own peculiar difficulty for observers, are Y Cyg, UZ Dra, AQ Peg, and AM Tau.

This past December Ray Berg and your committee chairman initiated a program to visually observe a few Hipparcos stars suspected of being eclipsing binaries. By April we felt that we had enough data to predict minima of CD Lyn, and that's when we called in the CCD observers. A detailed account of the team effort on this star appears below.

Following Hipparcos, other observing programs with automated features have been discovering variable stars by the thousands and listing them on the Internet. A good portion of these discoveries are eclipsing binaries. Many are well within the reach of visual observers, and nearly all are within the reach of amateur CCD observers.

Most of these new discoveries await determination or verification of their periods and establishment of elements to predict their minima. Accurate color photometry by CCD observers could reveal intimate details about these stars' characteristics. Some are bound to be very interesting. Got observing equipment, but can't find anything to do? There's a project in here somewhere.
--MEB

The Team Strikes Again: the CD Lyncis Campaign

The Hipparcos satellite found many new eclipsing binaries among stars brighter than 10th magnitude or so. Most of these are small-amplitude variables, but more than a dozen have minima of 0.5 magnitude or more. In the December 1999 Sky & Telescope, Roger Sinnott published finding charts for several of these eclipsing variables with unknown periods. One was CD Lyn, which Hipparcos found to have a range of 9.8-10.4 V and a possible period of 1.53 days.

Marv Baldwin and Ray Berg began visually monitoring some of these variables in December. They checked each star every hour or two on each clear night, but they didn't have any luck until February 24-25. That night, both Marv and Ray caught CD Lyn going into eclipse. Ray was clouded out as the variable reached minimum, but Marv was able to get more points on the rising branch. Then on March 4-5, in an 8-hour run, Marv recorded a second minimum.

The two visual minima were separated by 9.10 days, which was pretty close to six times 1.53 days. But the new times of minima pinned down phase 0.00. This permitted calculation of the phases of observations at maximum on other nights, which eliminated the 1.53-day period and some other possibilities. It was now likely that the period would be 9.10/n, with n = 1, 2, or 4.

The breakthrough came on the night of April 5-6, when Marv detected the beginning of a third minimum. He called Dan Kaiser to get CCD observations, but Dan couldn't observe because of high winds. So Dan e-mailed Arne Henden at the U.S. Naval Observatory's Flagstaff station. Arne just happened to be making CCD observations with the 1-meter reflector, and he swung the scope over to CD Lyn and got a good data set through the bottom of the minimum. The interval between the minima of March 4-5 and April 5-6 was 31.84 days. A few taps on a calculator showed that 31.84/7 = 9.10/2, so the best estimate of the period was now 4.55/n days.

The CCD network expanded to include Gary Billings in Calgary, Alberta; Gil Lubcke in Middleton, Wisconsin; and Bob Nelson at the College of New Caledonia in Prince George, British Columbia. Coordinated and encouraged by a steady exchange of e-mail, each observer attempted to catch CD Lyn at intervals of 4.55/n days after Arne's solid time of minimum. And 25 days later, Arne himself caught a minimum at the 0.5 phase of the 4.55 day period, establishing a new period of 2.27 days. Bob Nelson and Gary Billings added two more minima confirming the 2.27-day interval between minima.

Now the question was whether CD Lyn is a detached binary system with equally deep primary and secondary minima alternating every 2.27 days (whole period 4.55 days) or an Algol-type system with a deep primary minimum every 2.27 days and a shallow secondary minimum halfway between. CCD measures by Gary, Dan Kaiser and Gil Lubcke at the appropriate phases ruled out a secondary minimum greater than observational scatter between the deep minima every 2.27 days, indicating that CD Lyn is a detached system with virtually equal minima and a period of 4.55 days. "Virtually equal" means within 0.01 magnitude, according to Arne's high-precision observations at both primary and secondary minimum.

The Team's CCD minima timings only spanned an interval of 34 days or 7.5 orbital cycles. To refine the period value, a longer timeline was needed. So Your Editor and Tim Hager headed for the Harvard plate archives, where we estimated CD Lyn on more than 650 patrol plates and found 24 times of minimum dating back to 1901. This allowed the period to be calculated to seven significant decimal places.

As Lynx descended toward the evening twilight in May, the team wrote up its results, finding that CD Lyn is a detached system with equal minima, range 9.80-10.34 V, and P = 4.5494840 days. The paper was submitted to the Information Bulletin on Variable Stars and appeared in June as IBVS No. 4911 (www.konkoly.hu/cgi-bin/IBVS?4911).

This account has omitted a lot of the play-by-play action in the CD Lyn campaign, including the false clues provided by visual scatter and rogue points in the Hipparcos data that helped to confuse the first attempts to find the period. But that's part of the fun of solving these eclipsing binary puzzles. Want to get in on the action? More projects are underway or anticipated. Visual observers should contact Chairman Marv Baldwin, CCD observers should contact Dan Kaiser (dhkaiser@sprynet.com).
--DBW

Eclipsing Binary News & Notes

This discovery again proves the adage that you can never have too much data. After two months of observations, the authors had obtained a good spectroscopic orbital solution for the close pair. But they continued observing to reduce the formal errors in the solution. Then, when they had doubled the number of measures, they were puzzled to find that the errors had actually increased. Careful examination of the data showed systematic trends that proved to be the variation in the center-of-mass velocity caused by the third star. The mystery of the disappearing minima is explained by variation in the close binary's orbital inclination due to perturbations by the third component (or "secular regression of the nodes" in astrospeak).

The disappearance of eclipses could be explained by precession of the Algol system's orbit in the field of a third component, as in the case of SS Lac. The remaining 0.1 magnitude variation can't be due to tidally induced ellipticity of the B2 primary star, because that would produce two maxima at phases 0.25 and 0.75 of the 1.55-day orbit. Pulsation can be ruled out by the sheer improbability of the pulsation period just happening to equal the former Algol period. The most likely explanation that occurs to Your Editor is a reflection effect caused by the hot primary heating the facing side of the cool secondary, which would produce a maximum at phase 0.5 and a minimum at phase 0.0, just as the light curve shows.

And thirdly, the star Xi Tau has been known as a triple spectroscopic system for many years. Two sharp-lined stars orbit with a period of 7.15 days, and this pair revolves around a broad-lined B7 star every 145.1 days. Percy and Wilson (PASP 112, 846, 2000) call attention to the fact that the Hipparcos photometry of this system is constant except for several instances when the star is about 0.1 magnitude fainter than normal. The faint points occur at intervals of 7.15/2 days, indicating that Xi Tau is an eclipsing binary with similar primary and secondary minima.

Finally, a recent period decrease in HW Vir may indicate an unseen third companion according to Wood and Saffer (MNRAS 305, 820, 1999). HW Vir has a subdwarf B primary and an orbital period of only 2.8 hours. This eclipsing variable is a rare example of a post-common envelope binary.

According to these authors, "post-common envelope binaries are believed to have originated as much wider binaries in which the more massive star evolves into a red giant, fills its Roche lobe and transfers mass so rapidly on to the secondary star that it is entirely engulfed by the red giant (the common envelope phase). The secondary star then orbits inside the atmosphere of the red giant, losing angular momentum to the envelope as it does so. The secondary star therefore spirals inwards towards the core of the red giant and the envelope is ejected, leaving a subdwarf or white dwarf in a close orbit with the secondary star."

The authors present an O-C diagram (bottom of previous page) showing the recent period decrease, which can be represented by an abrupt change between two linear periods. However, it is unclear why a sudden mass transfer or ejection event would occur in this detached system. A period variation diagram showing the periods determined between pairs of minima plotted against time indicates that the period change could be a gradual decline. Such a gradual decline could be caused by magnetic braking, an extended common envelope remaining after the giant phase, or the presence of a third body in a noncircular orbit. Abrupt period change or gradual? "Further monitoring of the eclipse timings of HW Vir is necessary to distinguish between these two possibilities." Minima of this 10th magnitude eclipsing binary only last a fraction of an hour, and high-precision timings can be expressed to five significant decimal places of a day.

In BBSAG Bulletin 121, R. Diethelm presents a CCD light curve for NSV 25050 Cyg (above left), which is revealed as a faint Algol system, 13.5-14.8 (unfiltered), P = 1.41 days. E. Blattler contributes the first light curve of the W UMa-type star BI Vul (above right) which has a range (unfiltered) of 13.0-13.6 and the very short period of 0.25 day.

The early-type contact binary V606 Cen was classified as a supergiant system from objective prism spectra. Lorenz, Mayer, and Drechsel (Astron. Astrophys. 345, 531, 1999) present the first radial velocity curve and UBV light curves for this 9th-magnitude system. Their results for the masses, radii, and especially the surface gravities of both components indicate main-sequence stars with spectral types of B0-0.5V and B2-3V. This investigation adds V606 Cen to the short list of about a dozen early-type contact binaries for which accurate stellar parameters are known.

Several decades of photographic and visual minima timings of the 0.58-day eclipsing variable WZ Cyg show that the period is increasing. Rovithis et al. (Astron. Astrophys. 348, 184, 1999) report the first photoelectric study of the beta Lyr-type light curve. In visual light, the primary and secondary minima are 1.0 and 0.4 magnitude deep. There is some variability in the light curve from night to night and season to season. In addition, three brief flares of about 0.2 magnitude were recorded. Light curve analysis produced contact solutions. Nonetheless, the large difference between the primary and secondary minima shows that this binary is thermally decoupled and belongs to the interesting class of near-contact systems. Continued timings of minima are needed to monitor the increasing period.

The Information Bulletin on Variable Stars continues to serve as a fount of short papers on eclipsing binaries. This invaluable publication has moved to a new server at www.konkoly.hu. In IBVS 4819, Safar reports that GSC 3151-2126 is an Algol-type variable, P = 0.61 day, range 11.7 - 12.5 (unfiltered CCD). In IBVS 4836, Mkrtichian and Gamarova report that the primary component of R CMa is a delta Sct pulsating star with a period of 68 minutes and mean semi-amplitude of 4.4 mmag. Then in IBVS 4837, Gamarova, Mkrtichian, and Kusakin report that the primary component of AS Eri is another delta Sct star with P = 24 minutes and semi-amplitude 4.5 mmag.

In IBVS 4859, Sipahi, Keskin, and Yasarsoy present photoelectric light curves of the W UMa variable YY CrB, a bright Hipparcos discovery with a range of 8.6 - 9.1 V. In IBVS 4874, Berthold and Kroll report a period study of the Algol star RY CMi. The O-C diagram (below) can be interpreted as either two linear-period segments or a large and continuous decrease in period.

Featured Star: TU Herculis

The period of TU Her has been studied by several investigators. The most recent assessment by Qian Shengbang (Astron. J. 119, 901, 2000) confirmed a long-term decrease in the period at an overall rate of -2.92 x 10-6 days per year, within which trend two abrupt period changes also occurred.

The more or less steady decrease in the period of TU Her is unusual for an Algol system. Algol-type binaries are semidetached systems whose less-massive, evolved secondary stars fill their Roche lobes. Assuming conservative mass transfer from the lobe-filling secondary to the more massive primary component, the periods of Algol-type binaries should increase. Clearly, whatever is happening in the interactive TU Her binary system, it's not simple mass transfer from the secondary to primary component.

Shengbang's O-C diagram for TU Her from 1910 to 1997 is shown on the previous page, left. An earlier period study by Li et al. (Acta Astron. Sinica 7, 44, 1987) found that a smooth quadratic term fit all the times of minima then available (dashed line).

Newer data plotted after E = +2000 provide another fine demonstration of Williams' Law, which states, "If a period study produces a good linear or higher order fit to the observed times of minima of an eclipsing binary star, the period of that star will diverge from the calculated fit immediately following publication of the period study."

This sudden lengthening of the period and an earlier abrupt period decrease are illustrated dramatically in the second diagram (previous page, right), which shows the residuals from Shengbang's quadratic ephemeris (solid line in left diagram). The continuing overall decrease in the period of TU Her during the past 90 years plus the two abrupt period changes make this eclipsing binary a priority target for future timings of minima. Unfortunately, the period is very nearly equal to 2.25 days, so that observable minima occur only every 4P = 9.07 days. As a result, TU Her is often featured in the list of program stars that were not successfully observed during the previous year.