The Disappearing Act of KH 15D: Photometric Results from 1995 to 2004

We present results from the most recent (2002–2004) observing campaigns of the eclipsing system KH 15D, in addition to rereduced data obtained at Van Vleck Observatory (VVO) between 1995 and 2000. Phasing 9 yr of photometric data shows substantial evolution in the width and depth of the eclipses. The most recent data indicate that the eclipses are now approximately 24 days in length, or half the orbital period. These results are interpreted and discussed in the context of the recent models for this system put forward by Winn and coworkers and Chiang & Murray-Clay. A periodogram of the entire data set yields a highly significant peak at 48.37 ± 0.01 days, which is in accord with the spectroscopic period of 48.38 ± 0.01 days determined by Johnson and coworkers. Another significant peak, at 9.6 days, was found in the periodogram of the out-of-eclipse data at two different epochs. We interpret this as the rotation period of the visible star and argue that it may be tidally locked in pseudosynchronism with its orbital motion. If so, application of Hut's theory implies that the eccentricity of the orbit is e = 0.65 ± 0.01. Analysis of the Ultraviolet and Visual Echelle Spectrograph VLT spectra obtained by Hamilton and coworkers shows that the v sin i of the visible star in this system is 6.9 ± 0.3 km s-1. Using this value of v sin i and the measured rotation period of the star, we calculate the lower limit on the radius to be R = (1.3 ± 0.1) R⊙, which concurs with the value obtained by Hamilton and coworkers from its luminosity and effective temperature. Here we assume that i = 90°, since it is likely that the spin and orbital angular momenta vectors are nearly aligned. One unusually bright data point obtained in the 1995/1996 observing season at VVO is interpreted as the point in time when the currently hidden star B made its last appearance. Based on this datum, we show that star B is 0.46 ± 0.03 mag brighter than the currently visible star A, which is entirely consistent with the historical light curve. Finally, well-sampled VJ and IJ data obtained at the Cerro Tololo Inter-American Observatory Yale 1 m telescope during 2001/2002 show an entirely new feature: the system becomes bluer by a small but significant amount in very steady fashion as it enters eclipse and shows an analogous reddening as it emerges from eclipse. This suggests an extended zone of hot gas located close to but above the photosphere of the currently visible star. The persistence of the bluing of the light curve shows that its length scale is comparable to a stellar radius.