ZTF SN Ia DR2 follow-up: Characterization of subluminous Type Ia supernovae in the ZTF DR2 full sample

TL;DR

This study leverages the ZTF DR2 to assemble the largest homogeneous sample of subluminous Type Ia supernovae (124 events across 91bg-, 86G-, 04gs-, and 02es-like subtypes) and conducts a joint photometric–spectroscopic analysis. Through SALT2 and SNooPy light-curve fitting, Spextractor-based spectral measurements, spectral averaging, and host-galaxy contextualization, the authors map intrinsic colors, light-curve widths, and spectral features, revealing that subluminous SNe Ia are generally intrinsically redder and occur in more massive, redder environments, with 91bg-like events preferentially at larger galactocentric offsets. They show SALT2 color corrections can mask intrinsic color differences, while Ti II 4000 Å pEW provides a robust discriminator among subluminous subtypes, enabling refined classification and insights into progenitor populations. The results underscore potential systematics in using subluminous SNe Ia for cosmology if intrinsic color is not properly disentangled, and they offer a framework for subtype identification via spectral features in large SN samples. Overall, the work advances understanding of subluminous SN Ia diversity, environments, and implications for distance indicators.

Abstract

The Zwicky Transient Facility Data Release 2 (ZTF DR2) includes a total of 3,628 Type Ia supernovae (SNe~Ia), providing the largest and most complete sample of spectroscopically confirmed SNe~Ia at low redshift to date. In this paper, we present a photometric and spectroscopic analysis of 124 subluminous SNe~Ia, the largest sample of spectroscopically classified subluminous SNe~Ia observed with a single instrument, comprising 87 91bg-like, 12 86G-like, 18 04gs-like, and 7 02es-like events. We complement the published DR2 SALT2 light-curve parameters with new parameters obtained using template-based fits from SNooPy. Expansion velocities and pseudo-equivalent widths pEW of key spectral features are measured using Spextractor, and spectral averages are constructed for each subluminous subtype, binned by phase. We also analyze the host galaxy environments, both global and local, in terms of $g - z$ color, stellar mass, and directional light radius $d_{DLR}$. We find that all subluminous SNe~Ia (except the 02es-like subtype) are intrinsically red. This is evident by separating extrinsic from intrinsic color components. Since SALT2 is not trained on subluminous SNe~Ia, it compensates for their redder colors by inflating the $c$ parameter, thus extending the luminosity-width relation to negative values of x1. As expected, all subluminous SNe~Ia fall within the Cool region of the Branch et al. (2006) diagram, with the exception of 02es-like events, which show lower Si II 5972 pEW values. All subluminous subtypes tend to occur in more massive, redder host galaxies, and in the reddest local environments. Notably, 91bg- and 86G-like SNe~Ia explode at significantly larger normalized galactocentric distances. Finally, we identify the $pEW$ of the blended Ti II+Si II+Mg II absorption feature at 4300~A, along with s_BV, as robust and sufficient indicators for subclassifying subluminous SNe~Ia.

Figures (15)

• Figure 1: An example of SNooPy fit of ZTF18aaqfkqh light-curves. The three ZTF $gri$ bands do not have observations before peak magnitude, making the determination of the time of maximum very uncertain. Including $co$ ATLAS data allows for a combined fit that adequately constrain the time of maximum and the light-curve width. Note that the $ATgr$- and $ATri$-bands correspond to $c$- and $o$-ATLAS bands , respectively.
• Figure 2: Example spectrum of SN 2018aaz (ZTF18aabstmw) at 6 days before maximum light, highlighting the eight main spectral features. The corresponding pseudo-equivalent width (pEW) measurements are illustrated as blue-shaded regions.
• Figure 3: Demonstration of a measurement of a pseudo-equivalent width (pEW) for the SiII $\lambda$6355 feature with Spextractor. The rest-framed spectrum with its uncertainty is shown in red. The vertical dashed black lines determine the wavelength boundaries of the feature. A dashed red line represents the pseudo-continuum connecting both sides of the spectral feature, and the shaded blue area is the integral between the pseudo continuum and the binned spectrum within the defined wavelength boundaries, which defines the pEW.
• Figure 4: Spectral averages across six phase bins for all 103 91bg-like SNe Ia spectra in our sample. Key spectral features are marked with vertical lines at their minimum in the spectrum at maximum light, allowing their velocity evolution to be tracked across different phases.
• Figure 5: SN time of maximum light residuals comparison between SNooPy and SALT2 light-curve models. Blue points indicate differences of time of maxima when fitted to ZTF data only, and red dots indicate differences when ATLAS data is included in the SNooPy fits only.