Host stellar population properties and the observational selection function of type Ia supernovae.

摘要

Supernovae Ia are viable standard candles for measuring cosmological distances because of their enormous light output and similar intrinsic brightness. However, dispersion in intrinsic brightness casts doubt on the overall reliability of supernovae as cosmological distance indicators. Moreover, as shown in this thesis, the dependence of peak brightness on host galaxy properties significantly contributes to this dispersion. As a result, there is good reason to doubt that the nearby sample of supernovae Ia is identical to the distant samples, which occur in host galaxies that are billions of years younger. This study explores the validity of supernovae Ia as standard candles by examining regions of nearby galaxies that hosted supernovae and modeling their observational selection function. The approach is two-fold. First, photometry is performed on the stellar population environment of supernovae to characterize that region as a function of supernova type. Then, the observational selection function is simulated to determine the true supernovae production rate of the z < 0.1 redshift limit. We find that, on average, type Ia events occur in redder and older populations; underluminous supernovae Ia occur in regions that seem to be preferentially dusty, whereas normal Ia coming from the same galaxy type occur in a wide range of extinction environments. Furthermore, redder peak colors correspond to redder underlying population colors. This finding implies that dust extinction effects can cause systematic errors in the luminosity calibration of Ia events Finally, a single supernova rate does not adequately describe all supernovae Ia within z < 0.1. A rate of 0.25 SNu describes the population up to z < 0.03, and a much smaller rate, 0.1 SNu or less, describes supernovae past this distance. This finding indicates that observed supernova rates per galaxy remain biased by sample selection effects and that the intrinsic rate is likely uncertain by a factor of 2 to 3.