Observations of high redshift SNe are of interest for several reasons. First of all, one has through these a direct probe of the nucleosynthesis and star formation of the universe. In practice, there are several obstacles for a quantitative study of these issues. The fact that a large fraction of the star formation, and thus the SNe, may be hidden within optically thick dust can make it difficult to determine the total SFR and SNR accurately. This is certainly true for the optical bands, where we have found that the predicted total number of core collapse SNe with 1.5 is rather insensitive to the assumed star formation scenario, as long as the star formation is calculated to match the same observed luminosity densities, and the same extinction is assumed for the UV light from the galaxies and the SNe. Observations in the near-IR are less affected by this, and offers a clear advantage to the observations of the far-UV, as used for Lyman break objects. However, if a large fraction of the star formation occurs in highly obscured star burst galaxies, also the near-IR rates are severely affected. A further advantage of using SNe as star formation indicators is that they are insensitive to surface brightness selection effects. A complication when it comes to studying the nucleosynthesis is that the yields of the supernovae may vary with metallicity.
An important motivation for searches of SNe at high redshift is that one can from this type of observations learn something about the SNe themselves when observed in a different environment. In particular, differences in the fractions of the various core collapse subclasses, their spectra and luminosities may give new insight into the physics of the SNe and their progenitors.
The number of core collapse SNe that can be detected with NGST, with its expected limiting magnitude K´ = 31.4, should be 45 per field in a 104 s exposure, assuming a hierarchical star formation scenario. The mean redshift of these SNe is = 1.9. About one third of the SNe have 2. The high dust model results in total counts in the K´ band that are a factor 2 higher than in the hierarchical model. The estimated number of SNe with 2 in the K´ band for NGST is a factor 3 higher than in the hierarchical model. The model with flat SFR at high z, but with low extinction, result in a factor 2 higher number of SNe with 2, compared to the hierarchical model, for the NGST limit.
An important practical point is that in order to detect especially the Type IIP SNe at high z it is necessary to have a spacing between observations of 100 days for ground based telescopes, and 1 year for deep observations with NGST. Shorter time intervals do not allow for the luminosity of the SNe to decrease by an amount necessary for detection.
When it comes to the observed rates of Type Ia SNe, we find that these are highly sensitive to the star formation modeling. This is due to the fact that the Type Ia's are less linked to the environment where their progenitors were formed. The uncertainty in the life-time of the progenitors, combined with the sensitivity of the Type Ia rates to the onset of star formation in the models with a flat SFR at high z, contributes to the difficulty with using Type Ia counts as probes for either different star formation scenarios or progenitor models. This is further hampered by the fact that even the local rate is highly uncertain, and that this propagates to other redshifts through the normalization of the rates at z = 0. Therefore, accurate measurements of the Type Ia rates at low z are most desired.
Precise measurements of the Type Ia rates at 1 could constrain the parameters to some extent. For example, in a given cosmology there is a high redshift cutoff in the Ia rates at an epoch that depends strongly on , but is less dependent on the star formation scenario. In agreement with previous studies we find that counts of Type Ia SNe can be used as cosmological probes. This does, however, require that both the SFR and the unknown life time of the Type Ia progenitors can be determined independently.
We predict the number of simultaneously detectable Type Ia SNe per NGST field to be , depending on progenitor model and star formation scenario. Additional uncertainties widen this range even more. Of the simultaneously observable Type Ia SNe, about 5% are on the rising part of the light curve. For a ground based telescopes with limiting magnitude IAB 27 we predict Type Ia's per square degree of which 30% are on the rise of the light curve.
A major technical aspect of our work is that we have tried to incorporate as much knowledge as possible about the theoretical and observational properties of the different classes of SNe. In particular, we have found that the spectral evolution is important for the magnitudes in the different bands. A striking example is the sensitivity to the UV cutoff, which for most SNe dominate the evolution in the optical bands. We have also found that a proper treatment of the light curve can change the predicted rates by factors of three or larger. An important source of uncertainty in these estimates are the local frequencies of SNe of different classes, as well as their distribution in luminosity. More extensive surveys with well defined selection criteria are therefore of highest priority for reliable predictions at high redshifts.
© European Southern Observatory (ESO) 1999
Online publication: October 4, 1999