The Type Ia supernova 2021hem: A 2003fg-like event in an apparently hostless environment

We report observations of a Type Ia supernova (SN Ia) 2021hem that was discovered within 48 hours of last nondetection and is located in an apparently hostless environment. With a peak absolute $B$-band magnitude of $M_{B,\mathrm{max}} = -19.96 \pm 0.29$ mag, SN 2021hem lies at the luminous end of the SNe~Ia distribution. Its near-infrared and $i$-band light curves lack the secondary maximum, which is otherwise ubiquitous to normal and 1991T-like SNe~Ia. Instead, these properties cause SN~2021hem to closely resemble 2003fg-like events. The slowly evolving light curves (characterized by $\Delta m_{15}(B) = 1.02 \pm 0.02$ mag; $s_{BV} = 0.94 \pm 0.05$) and the earliest spectrum showing C II $\lambda 6580$ and $\lambda 7235$ absorption lines further support this classification. Other spectroscopic features, including Si~II line diagnostics, resemble those of normal SNe Ia. A fit of a fireball model to the early-time light curves yields a time of first light of $t_{\mathrm{first}} = -16.43^{+0.45}$ -0.38 days relative to $B$-band maximum. The first photometric detection occurs $1.51^{+0.45}$ -0.38 days before the onset of fireball-like flux rise. This early emission, together with the intrinsic $(g - r)0$ color, is inconsistent with circumstellar or companion interaction. Instead, shallow $^{56}$Ni mixing or an asymmetric $^{56}$Ni distribution offers a plausible explanation for the delayed onset of the fireball flux rise, while a double-detonation scenario with a thin helium shell remains a less likely alternative. Notably, SN 2021hem represents the fifth known 2003fg-like SN that has early-time activity or excess flux emission. The estimated mass of radioactive $^{56}$Ni synthesized in SN 2021hem is $1.00 \pm 0.09 \mathrm{M}{\odot}$. Deep GTC imaging obtained 2.5 years after the explosion, with an estimated limiting magnitude of $m_{\mathrm{lim},r} = 24.4$ mag and a surface-brightness limit of $\mu_{\mathrm{lim},r} = 26.3$ mag arcsec$^{-2}$, revealed no coincident host. Most faint dwarf and ultradiffuse galaxies (UDGs) are therefore ruled out. Alternatively, if the nearest plausible AGN host galaxy located at a projected distance of 104 kpc is assumed, the progenitor would need to be a hypervelocity star ejected at $\approx 2200$ km s$^{-1}$ from the host by AGN interaction. A faint diffuse feature $\approx 6$ kpc from the SN site has also been detected in the GTC image, and its surface brightness is within the limits of UDGs. It is unclear whether it is a galaxy and is associated with SN 2021hem, however. Based on its large normalized directional light distance ($d_{\mathrm{DLR}} \approx 3$--$4$) from the SN and its unusual elongation, the probability that this is the candidate host galaxy of SN 2021hem is low. These results identify SN 2021hem as one of the strongest candidates for a hostless SN Ia and underscore the diversity of luminous slowly evolving 2003fg-like explosions and the wide range of environments in which they may occur.