Paper detail

Phase resolved spectroscopy and Kepler photometry of the ultracompact AM CVn binary SDSS J190817.07+394036.4

{\it Kepler} satellite photometry and phase-resolved spectroscopy of the ultracompact AM CVn type binary SDSS J190817.07+394036.4 are presented. The average spectra reveal a variety of weak metal lines of different species, including silicon, sulphur and magnesium as well as many lines of nitrogen, beside the strong absorption lines of neutral helium. The phase-folded spectra and the Doppler tomograms reveal an S-wave in emission in the core of the He I 4471 Å\,absorption line at a period of $P_{\rm orb}=1085.7\pm2.8$\,sec identifying this as the orbital period of the system. The Si II, Mg II and the core of some He I lines show an S-wave in absorption with a phase offset of $170\pm15^\circ$ compared to the S-wave in emission. The N II, Si III and some helium lines do not show any phase variability at all. The spectroscopic orbital period is in excellent agreement with a period at $P_{\rm orb}=1085.108(9)$\,sec detected in the three year {\it Kepler} lightcurve. A Fourier analysis of the Q6 to Q17 short cadence data obtained by {\it Kepler} revealed a large number of frequencies above the noise level where the majority shows a large variability in frequency and amplitude. In an O-C analysis we measured a $\vert\dot{P}\vert\sim1.0\,$x$\,10^{-8}\,$s\,s$^{-1}$ for some of the strongest variations and set a limit for the orbital period to be $\vert\dot{P}\vert<10^{-10}$s\,s$^{-1}$. The shape of the phase folded lightcurve on the orbital period indicates the motion of the bright spot. Models of the system were constructed to see whether the phases of the radial velocity curves and the lightcurve variation can be combined to a coherent picture. However, from the measured phases neither the absorption nor the emission can be explained to originate in the bright spot.