White dwarf masses in cataclysmic variables

Abstract: The white dwarf (WD) mass distribution of cataclysmic variables (CVs) has recently been found to dramatically disagree with the predictions of the standard CV formation model. The high mean WD mass among CVs is not imprinted in the currently observed sample of CV progenitors and cannot be attributed to selection effects. Two possibilities have been put forward: either the WD grows in mass during CV evolution, or in a significant fraction of cases, CV formation is preceded by a (short) phase of thermal time-scale mass transfer (TTMT) in which the WD gains a sufficient amount of mass. We investigate if either of these two scenarios can bring theoretical predictions and observations into agreement. We employed binary population synthesis models to simulate the present intrinsic CV population. We incorporated aspects specific to CV evolution such as an appropriate mass-radius relation of the donor star and a more detailed prescription for the critical mass ratio for dynamically unstable mass transfer. We also implemented a previously suggested wind from the surface of the WD during TTMT and tested the idea of WD mass growth during the CV phase by arbitrarily changing the accretion efficiency. We compare the model predictions with the characteristics of CVs derived from observed samples. We find that mass growth of the WDs in CVs fails to reproduce the observed WD mass distribution. In the case of TTMT, we are able to produce a large number of massive WDs if we assume significant mass loss from the surface of the WD during the TTMT phase. However, the model still produces too many CVs with helium WDs. Moreover, the donor stars are evolved in many of these post-TTMT CVs, which contradicts the observations. We conclude that in our current framework of CV evolution neither TTMT nor WD mass growth can fully explain either the observed WD mass or the period distribution in CVs.