Pre-maximum spectroscopic diversity of hydrogen-poor superluminous supernovae

We search for the reasons behind the spectroscopic diversity of hydrogen-poor superluminous supernovae (SLSNe-I) in the pre-maximum phase. Our analysis is a continuation of the paper of \citet{ktr21}, who disclosed two new subtypes of SLSNe-I characterized by the presence/absence of a W-shaped absorption feature in their pre-maximum spectra between 4000 and 5000 \AA (called Type~W and Type~15bn, respectively). However, the physical cause of this bimodality is still uncertain. Here we present pre-maximum spectral synthesis of 27 SLSNe-I with special attention to the photospheric temperature ($T_{ \rm phot}$) and velocity ($v_{ \rm phot}$) evolution. We find that a $T_{\rm phot}$ limit of 12000~K separates the Type~W and Type~15bn SLSNe-I: Type~W objects tend to show $T_{\rm phot}\geq$12000~K, while Type~15bn ones have $T_{\rm phot} \leq$12000~K. This is consistent with the chemical composition of the studied objects. Another difference between these groups may be found in their ejecta geometry: Type W SLSNe-I may show null-polarization, implying spherical symmetry, while the polarization of Type 15bn objects may increase in time. This suggests a two-component model with a spherical outer carbon-oxygen layer and an asymmetric inner layer containing heavier ions. The two subgroups may have different light curve evolution as well, since 6 Type~W objects show early bumps, unlike Type 15bn SLSNe-I. This feature, however, needs further study, as it is based on only a few objects at present.