The Eclipsing Binary Di Herculis: One Mystery Solved, But Another Takes Its Place

The 8^th-mag eclipsing binary DI Herculis has perplexed scientists for the past few decades due to its anomalously slow apsidal motion rate. DI Her consists of two main-sequence stars (B5V, B6V), with P(orb) = 10.55 days, and eccentricity(e= 0.489). Since the apsidal motion is dominated by General Relativity, the system is one of the few tests available for verifying the theory. Combining the expected classical (1.93°/100 yr) and relativistic (2.34°/100 yr) effects, the predicted apsidal motion rate is 4.27°/100 yr. Our recent determination of the apsidal motion yields 1.33°+/-0.25 /100 yr, based on eclipse timings from 1936-2008.

Recently, Albrecht et al (2009, Nature 461) have apparently solved the apsidal motion anomaly of DI Her, finding that the axes of both stars are significantly inclined from the normal to the orbital plane. This was determined from the radial velocity curves and observing the Rossiter-McLaughlin effect during primary and secondary eclipses. Having significantly misaligned axes of rotation produces a perturbation that greatly reduces the classical apsidal motion effect, thus explaining the observed small apsidal motion rate. Even though this discovery apparently solves the problem, it raises new questions as to how the axes are so tilted. Additionally, tilted axes are expected to contribute to other orbital effects, such as changes in orbital inclination, which have not yet observed from the apparent constancy in eclipse depths over time. We have also searched for evidence of small periodic oscillations in the eclipse timings and found no evidence of a light travel time effect arising from a possible tertiary component. Further, we find evidence that the projected rotation axes of the stars may be precessing, since it appears that the value of V(rot)sini has increased over the past 30 years. This research was supported by NSF/RUI Grants AST05-07536/42.