2012A&A...539A..94B

CP Pup (Nova Pup 1942) showed outburst and quiescent characteristics indicating a very massive white dwarf, yet the standard spectroscopic dynamical analysis assuming an accretion disk yields an extremely low value for the white dwarf mass. However, some physical parameters and the accretion geometry are still poorly known. The nova was spectroscopically monitored between 1988 and 1996. We analyzed the whole data set in order to re-determine the spectroscopic period and examine its stability. We also looked for chemical anomalies in the spectrum. We obtained the radial velocity curves for the hydrogen and helium lines from our last better quality 1996 run. The mean 1996 spectrum yields information on the chemical composition of the binary. We also searched the mean period using the multi-year data set. From the radial velocities of our complete data set we derive the most probable mean spectroscopic period and tentatively suggest revised ephemeris. However, we demonstrate that the period is intrinsically unstable. We show that a standard accretion disk model does not explain all the spectroscopic features observed nor the variability of the spectroscopic period. We suggest that only interpreting the system as a slightly asynchronous polar would fit the data. The mean optical spectrum of CP Pup shows also an enhanced carbon abundance. Non solar abundances in the accreted material are unexpected and interesting, confirming that the nature of the secondaries of old novae should be studied more in detail. In fact, in CP Pup, as in other novae, the enhanced carbon is an important clue to the pre-outburst evolution, implying that the secondary was heavily polluted with carbon and helium during the common envelope phase of the pre-cataclysmic binary by a relatively massive primary that filled its Roche lobe during the third dredge up on the asymptotic giant branch.