17191

Type Ia supernovae (SNe Ia) have long been used as standard candles to measure extragalactic distances and cosmological parameters. SNe Ia are portrayed as a homogeneous class comma but there are actually several subtypes. Even the d_commanormald_comma SNe Ia used for cosmology may not be homogeneous. Recently comma the optical light curves of SNe Ia have been shown to slow down comma relative to their earlier decline rate comma at >800 days after explosion. Moreover comma more luminous SNe Ia may slow down faster than less luminous objects. This correlation comma which recalls the peak-light stretch-luminosity relation used to standardize SNe Ia comma also hints at the existence of several production channels for normal SNe Ia comma as no single explosion model can produce the full range of late-time light curves. But this new correlation is based on just 6 objects. We ask for 67 WFC3-UVIS orbits (with the F438W comma F555W comma and F814W filters) spread over Cycles 30-32 to observe 12 SNe Ia when they are 600-1200 days old. By tripling the sample of late-time SNe Ia comma we will prove the existence of the new stretch-luminosity correlation at a significance of >5-sigma. The impact on SN cosmology will be twofold. First comma a new stretch-luminosity correlation could further standardize SNe Ia and reduce systematic uncertainties. Second comma strong evidence for the existence of multiple production channels for normal SNe Ia would force cosmologists to revisit their use of SNe Ia as a monolithic class. This experiment is time critical. Every year comma on average comma <5 SNe Ia can be used for this experiment. We are lucky to have 12 targets to work with this year this chance may not come again during HST.s remaining lifetime.