The bolometric light curve modeling of 98 Type I superluminous supernovae using the magnetar- and the circumstellar interaction models reveals surprisingly high ejecta masses

We present the bolometric light curve modeling of 98 hydrogen-poor superluminous supernovae (SLSNe-I) using three types of power inputs: the magnetar model and two kinds of circumstellar interaction models, applying the constant density and the steady wind scenario. The quasi-bolometric luminosities of the objects were calculated from the ZTF g- and r-band data using the methodology of \citet{chen23b}, and then they were modeled with the Minim code. It was found that the light curves of 45 SLSNe-I can be fitted equally well with both the magnetar and the CSM models, 14 objects prefer the magnetar model and 39 SLSNe-I favor the CSM model. The magnetar modeling yielded a mean spin period of $P~=~4.1 \pm 0.20$ ms and a magnetic field of $B~=~5.65 \pm 0.43 \cdot 10^{14}$ G, consistently with the literature. However, the ejected mass was estimated to be significantly larger compared to previous studies presenting either multi-color light curve modeling with MOSFiT or bolometric light curve modeling: we obtained a mean value and standard error of 34.26 and 4.67 $M_\odot$, respectively. The circumstellar interaction models resulted in even larger ejecta masses with a mean and standard error of 116.82 and 5.97 $M_\odot$ for the constant density model, and 105.99 and 4.50 $M_\odot$ for the steady wind model. Although the ejected mass depends strongly on the electron scattering opacity (assumed to be $\kappa~=~$0.2 in this work) and the ejecta velocity, which were estimated to be globally larger compared to earlier studies, our results suggest that SLSNe-I are indeed the explosions of the most massive stars.
Comment: accepted for publication in ApJ, 23 pages, 29 figures, supplementary at https://doi.org/10.5281/zenodo.14591928