Neutrinos play a crucial role in the collapse and explosion of massive stars, governing the infall dynamics of the stellar core, triggering and fueling the explosion and driving the cooling and deleptonization of the newly formed neutron star. Due to their role neutrinos carry information from the heart of the explosion and, due to their weakly interacting nature, offer the only direct probe of the dynamics and thermodynamics at the center of a supernova. In this paper, we review the present status of modelling the neutrino physics and signal formation in collapsing and exploding stars. We assess the capability of current and planned large underground neutrino detectors to yield faithful information of the time and flavor dependent neutrino signal from a future Galactic supernova. We show how the observable neutrino burst would provide a benchmark for fundamental supernova physics with unprecedented richness of detail. Exploiting the treasure of the measured neutrino events requires a careful discrimination of source-generated properties from signal features that originate on the way to the detector. As for the latter, we discuss self-induced flavor conversions associated with neutrino-neutrino interactions that occur in the deepest stellar regions; matter effects that modify the pattern of flavor conversions in the dynamical stellar envelope; neutrino-oscillation signatures that result from structural features associated with the shock-wave propagation as well as turbulent mass motions in post-shock layers. Finally, we highlight our current understanding of the formation of the diffuse supernova neutrino background and we analyse the perspectives for a detection of this relic signal that integrates the contributions from all past core-collapse supernovae in the Universe.

The rise time of a Galactic supernova (SN) bar-nue lightcurve, observable at a high-statistics experiment such as the IceCube Cherenkov detector, can provide a diagnostic tool for the neutrino mass hierarchy at "large" 1-3 leptonic mixing angle theta_13. Thanks to the combination of matter suppression of collective effects at early postbounce times on one hand and the presence of the ordinary Mikheyev-Smirnov-Wolfenstein effect in the outer layers of the SN on the other hand, a sufficiently fast rise time on O(100) ms scale is indicative of an inverted mass hierarchy. We investigate results from an extensive set of stellar core-collapse simulations, providing a first exploration of the astrophysical robustness of these features. We find that for all the models analyzed (sharing the same weak interaction microphysics) the rise times for the same hierarchy are similar not only qualitatively, but also quantitatively, with the signals for the two classes of hierarchies significantly separated. We show via Monte Carlo simulations that the two cases should be distinguishable at IceCube for SNe at a typical Galactic distance 99% of the times. Finally, a preliminary survey seems to show that the faster rise time for inverted hierarchy as compared to normal hierarchy is a qualitatively robust feature predicted by several simulation groups. Since the viability of this signature ultimately depends on the quantitative assessment of theoretical/numerical uncertainties, our results motivate an extensive campaign of comparison of different code predictions at early accretion times with implementation of microphysics of comparable sophistication, including effects such like nucleon recoils in weak interactions.