IMPLICATIONS OF THE DELAYED 2013 OUTBURST OF ESO 243-49 HLX-1

After showing four quasi-periodic outbursts spaced by similar to 1 yr from 2009 to 2012, the hyper luminous X-ray source ESO 243-49 HLX-1, currently the best intermediate mass black hole (IMBH) candidate, showed an outburst in 2013 delayed by more than a month. In Lasota et al., we proposed that the X-ray light curve is the result of enhanced mass transfer episodes at periapsis from a donor star orbiting the IMBH in a highly eccentric orbit. In this scenario, the delay can be explained only if the orbital parameters can change suddenly from orbit to orbit. To investigate this, we ran Newtonian smooth particle hydrodynamical simulations starting with an incoming donor approaching an IMBH on a parabolic orbit. We survey a large parameter space by varying the star-to-black hole mass ratio (10(-5)-10(-3)) and the periapsis separation r(p) from 2.2 to 2.7r(t) with r(t), the tidal radius. To model the donor, we choose several polytropes (Gamma = 5/2, n = 3/2, Gamma = 3/2, n = 2, Gamma = 5/3, n = 2, and G = 5/3, n = 3). Once the system is formed, the orbital period decreases until reaching a minimum that may be shallow. Then, the period tends to increase over several periapsis passages due to tidal effects and increasing mass transfer, leading ultimately to the ejection of the donor. We show that the development of stochastic fluctuations inside the donor by adding or removing orbital energy from the system could lead to sudden changes in the orbital period from orbit to orbit with the appropriate order of magnitude to that which has been observed for HLX-1. We also show that given the constraints on the black hole (BH) mass (M-BH > 10(4) M-circle dot) and assuming that the HLX-1 system is currently near a minimum in period of similar to 1 yr, the donor has to be a white dwarf or a stripped giant core. We predict that if HLX-1 is indeed emerging from a minimum in orbital period, then the period would generally increase with each passage, although substantial stochastic fluctuations can be superposed on this trend.