2010ApJ...712L.143S

The total masses ejected during classical nova (CN) eruptions are needed to answer two questions with broad astrophysical implications: can accreting white dwarfs be "pushed over" the Chandrasekhar mass limit to yield type Ia supernovae? Are ultra-luminous red variables a new kind of astrophysical phenomenon, or merely extreme classical novae? We review the methods used to determine nova ejecta masses. Except for the unique case of BT Mon (nova 1939), all nova ejecta mass determinations depend on untested assumptions and multi-parameter modeling. The remarkably simple assumption of equipartition between kinetic and radiated energy (E_kin and E_rad, respectively) in nova ejecta has been invoked as a way around this conundrum for the ultra-luminous red variable in M31. The deduced mass is far larger than that produced by any CN model. Our nova eruption simulations show that radiation and kinetic energy in nova ejecta are very far from being in energy equipartition, with variations of 4 orders of magnitude in the ratio E_kin/E_rad being commonplace. The assumption of equipartition must not be used to deduce nova ejecta masses; any such "determinations" can be overestimates by a factor of up to 10,000. We data-mined our extensive series of nova simulations to search for correlations that could yield nova ejecta masses. Remarkably, the mass ejected during a nova eruption is dependent only on (and is directly proportional to) E _rad. If we measure the distance to an erupting nova and its bolometric light curve, then E_rad and hence the mass ejected can be directly measured.