Despite a century of intense theoretical and observational investigation, the origin of Galactic Cosmic Rays (GCRs) remains poorly understood at present. Supernova remnants remain the most popular site for the acceleration of GCR particles (mostly protons), but the various candidate physical processes have not been sufficiently elucidated yet. The prominent mechanism, diffusive shock acceleration, involves several highly interconnected processes, including: the propagation of the supernova shock waves in a dynamically evolving environment (shaped by the physics of the pre-supernova star and, perhaps, of nearby stars as well), the acceleration of a fraction of the particles/nuclei encountered by the shock waves, the back-reaction of the accelerated particles on the system, the amplification of the magnetic field due to various instabilities (and the dynamical reaction of the amplified magnetic field on the plasma), etc. As a result, the system is highly non-linear and requires various techniques (analytical, numerical, Monte Carlo) to study its behavior. Important clues to the origin of GCRs are provided by their energy spectra and source composition, as well as by the electromagnetic signatures they emit in various wavelengths, both in their putative acceleration sites (SNRs) and throughout the Galactic interstellar medium where they propagate.

The characteristic signature of accelerated protons (gamma-rays from the decay of pions, produced through energetic proton collisions) was for a long time difficult to distinguish from emission of energetic electrons (bremsstrahlung or inverse Compton scattering). Recently, the pion-decay feature was detected with the Fermi Large Area Telescope in the gamma-ray spectra of two SNRs, IC 443 and W44,: both of these are remnants of core-collapse supernovae and IC443 is a member of the OB association GEM OB1. This detection provides direct evidence that cosmic-ray protons are indeed accelerated in SNRs. On the other hand, the GCR source composition presents strong similarities to the standard (cosmic) composition, but with noticeable differences. The most important one, from the nucleosynthesis point of view, is the high isotopic ratio of 22Ne/20Ne (5 times higher in GCRs than in the Sun), which requires a contribution of material accelerated from the winds of Wolf-Rayet (WR) stars, enriched in 22Ne. Other differences include the overabundances of some volatile or semi-volatile elements (like N, C and O) and the large overabundance of refractories (like Fe-peak elements) which strongly suggests sputtering of inter- (or circum-) stellar grains as a possible origin.

There is no model of GCR acceleration accounting in a self-consistent way for the detailed GCR source composition. We propose a team project to undertake a thorough investigation of the problem of GCR acceleration and composition in a realistic astrophysical setting, considering each one of the key factors:

• properties of the supernova progenitor stars (including the density, velocity and composition of their winds); we will consider SN from massive star progenitors – which account for 80% of all SN in the Galaxy – and SNIa.

• properties of the circumstellar medium of the supernovae (as shaped by the winds of the progenitor stars and/or their companions), including the composition of gas and dust,

• propagation of the supernova shock wave in the surrounding medium,

• particle acceleration in that environment (forward and reverse shocks, interactions with magnetic fields, grain sputtering and acceleration of various species),

• resulting observational signatures (photon and particle spectra, particle composition) taking properly into account an average over the massive star and SN population of the Galaxy,

• comparison with existing observations, analysis of implications, suggestions for forthcoming experiments.