2008ApJ...684...68F

We present a large-scale effort of creating composite spectra of high-redshift SNe Ia and comparing them to low-redshift counterparts in an attempt to understand possible cosmic evolution of SNe Ia, which has major implications for studies of dark energy. Through the ESSENCE project, we have obtained 107 spectra of 88 high-redshift SNe Ia with excellent light-curve information. In addition, we have obtained 397 spectra of low-redshift SNe Ia through a multiple-decade effort at the Lick and Keck Observatories, and we have used 45 UV spectra obtained by HST and IUE. The low-redshift spectra act as a control sample when comparing to the ESSENCE spectra. In all instances, the ESSENCE and Lick composite spectra appear very similar. The addition of galaxy light to the Lick composite spectra allows an excellent match of the overall SED with the ESSENCE composite spectra, indicating that the high-redshift SNe are more contaminated with host galaxy light than their low-redshift counterparts. This is caused by observing objects at all redshifts with similar angular slit widths, which corresponds to different projected physical distances. After correcting for the galaxy light contamination, a few marginally significant differences in the spectra remain. We have estimated the systematic errors when using current spectral templates for K-corrections to be ∼0.02 mag. The variance in the composite spectra gives an estimate of the intrinsic variance in low-redshift maximum light SN spectra of ∼3% relative flux in the optical and growing toward the UV. The difference between the maximum light low- and high-redshift spectra constrains the evolution of SN spectral features between our samples to be <10% relative flux in the rest-frame optical. Currently, galaxy contamination and the small samples of rest-frame UV spectra at low and high redshifts are the limiting factors for future studies.