2002A&A...392..541S

We present an analysis of the X-ray light curves of the magnetic cataclysmic variable DP Leo using recently performed XMM-Newton EPIC and archival published and unpublished ROSAT PSPC observations. We combine the timings of the X-ray eclipses with timings derived from archival HST-observations and new optical observations with the photon counting OPTIMA camera. We determine the eclipse length at X-ray wavelengths to be 235±5s, slightly longer than at ultra-violet wavelengths, where it lasts 225s. A new orbital ephemeris is derived which connects the more than 120000 binary cycles covered since 1979. It has a highly significant quadratic term, implying an orbital period change of {dot}(P)=-4.4x10^–12s/s, two orders of magnitude larger than being compatible with braking by gravitational radiation only. Over the last twenty years, the optical and X-ray bright phases display a continuous shift with respect to the eclipse center by ∼2.1°/yr. Over the last 8.5 years the shift of the X-ray bright phase is ∼2.5°/yr. We interpret this as evidence of an asynchronously rotating white dwarf although synchronization oscillations cannot be ruled out completely. If the observed phase shift continues, a fundamental rearrangement of the accretion geometry must occur on a time-scale of some ten years. Applying model atmosphere spectra to optical/UV eclipse light curves, we determine the temperature and mass of the white dwarf, the temperature and size of the optical/UV emitting spot and the distance to DP Leo to be T_wd=13500K, M_wd≃0.6M_☉, T_spot=32000K, A_spot≃0.1A_wd, and D=400pc, respectively. The implied inclination and mass ratio are i=79.5° and Q=M_wd/M₂=6.7. DP Leo is marginally detected at eclipse phase in X-rays. The upper limit eclipse flux is consistent with an origin on the late-type secondary, L_X≃2.5x10²⁹ergs/s (0.20-7.55keV), at a distance of 400pc.