ZTF-SEDm Type Ia supernova sample for Twins Embedding spectrophotometric standardisation

TL;DR

This work extends spectrophotometric SN Ia standardisation to the ZTF DR2 dataset by constructing a flux-calibrated ZTF-SEDm sample and applying the Twins Embedding framework. The authors implement a three-step TE pipeline—phase correction, Read Between The Lines color de-reddening, and non-linear manifold-based parameterisation—on 783 ZTF spectra, achieving a phase-corrected dispersion of $0.01$ mag in $g$ and a RBTL-driven Hubble residual scatter of $0.153$ mag (nMAD), with an upper limit of $0.129$ mag when accounting redshift and calibration errors. Compared to the SNfactory baseline, the ZTF-based RBTL standardisation shows reduced astrophysical bias in host steps and demonstrates that full TE gains require higher spectral quality; nonetheless, a sizeable, flux-calibrated subset (1897 spectra from 1607 SNe Ia) is released, enabling spectrophotometric standardisation even with limited data quality. The results validate TE applicability to large, heterogeneous spectroscopic surveys and highlight the importance of robust flux calibration and spectral SNR for fully leveraging non-linear TE components. This work informs the design of future spectroscopic surveys and the interpretation of SN Ia distances in cosmology, particularly in the presence of host contamination and calibration systematics.

Abstract

This paper has two aims: the first one is to build a large homogeneous spectrophotometric Type Ia supernova (SN Ia) sample, using 3069 spectra from the second Zwicky Transient Facility data release (ZTF DR2). Using this sample we reproduce, as the second objective of the paper, the Twins Embedding (TE) spectrophotometric standardisation method, which led to an exceptionally low value of 0.073 mag for the intrinsic scatter. We improve the flux-calibration accuracy of the SEDm SN Ia spectral sample using the ZTF photometric data, which are calibrated at the percent level. We then apply the three steps of the TE parameterisation to a subset of 783 ZTF SN spectra near maximum light, and analyse the resulting standardisation methods. The precision of the phase correction model, which is the first step of the TE, is estimated at 0.01 mag in g band, using ZTF data. Despite the challenge posed by the ZTF spectrum extraction pipeline, we apply a first standardisation in color based on the second step of the TE, the Read Between The Lines (RBTL). When considering the scatter due to the redshift error and the flux calibration error, we estimate a 0.129 mag Hubble residual scatter for this ZTF sample as an upper limit. As expected from the low spectral quality, the final TE standardisation based on three non-linear parameters did not improve the overall dispersion. We release 1897 flux calibrated spectra of 1607 SNe Ia with an estimated photometric accuracy of 0.07 mag. We further demonstrate the ability to apply a spectrophotometric standardisation with limited quality spectra. The RBTL standardisation is more efficient than that of SALT with one less parameter, and the resulting host steps are consistent with zero, making it less prone to astrophysical bias. For future spectroscopic surveys, a better spectral quality would enable the full TE standardisation to be computed. (Abridged)

Figures (21)

• Figure 1: SN ZTF20aayvubx spectrum at phase $0.01$ day is shown in blue, and after flux calibration in red. The camera's filters bandpasses are displayed in orange. The blue dots are the spectrum integrated through the filters, the green ones are the LC values. The offsets are 0.780, 0.708, 0.805 mag in $g$, $r$ and $i$ respectively. Second order polynomials, passing through the three photometric points, are shown in their respective colors.
• Figure 2: Magnitude offsets between the 1897 ZTF integrated spectra and the associated LC values, before calibration, in $g$, $r$ and $i$ filters. STD are shown in the legend. The dashed black lines indicate the LC expectations, and the 2-$\sigma$ density levels of the 2d histograms are shown in dark red.
• Figure 3: Top plot shows the precision (RMS) of the flux calibration as a function of wavelength, for 6 standard stars in light purple and the quadratic average in red. It shows the 0.03 mag and 0.012 mag limits by dashed lines. Middle plot shows the dispersion (nMAD) of the residuals relative to a SALT 2.4 spectrum model given LC data, for 1897 ZTF spectra, and 583 SNfactory spectra. It shows as well in light green the quadratic addition of the nMAD of SNfactory and 0.073 mag. The wavelengths delimiting a good precision are shown by dashed lines, at 4500 Å and 7000 Å. Bottom plot shows the mean SALT models residuals dispersion, binned in SNR, for the ZTF (75 SNe per bin). A 0.04 mag floor is shown by red line.
• Figure 4: Top panels show the calibration coefficients $a_0$, $a_1$ and $a_2$ distributions, for 1897 spectra, and their mean and STD are indicated in the legends. Bottom panels show the $d_{DLR}$ against the calibration coefficients. Dark blue datapoints show the coefficients mean and STD binned on 10 $d_{DLR}$ values. Vertical dashed black lines indicate the values corresponding to no correction, and horizontal lines correspond to the $d_{\text{DLR}}=1$ limit.
• Figure 5: Distributions of SN redshifts, phases nearest maximum, and SNR defined in the range 6700--7300 Å, for the flux-calibrated sample of ZTF SNe Ia after DR2 cuts (by blue dashed line), the sub-sample for TE application (in blue), and the SNfactory sample (in red).