Non linear cosmic ray transport and cosmic ray driven galactic winds

The bulk of Cosmic Rays (CRs) reaching our planet is likely of Galactic origin and is thought to be accelerated in sources located in the Galactic disk (mainly Supernova remnants (SNRs)). The energy density of Galactic CRs, ∼ 1 eV/cm3, can be accounted for if one assumes that 3-10% of the mechanical energy injected by SNe in the Galaxy is channeled into accelerated particles. The observed large residence time of CRs in the Galaxy (compared to the time required for ballistic propagation of relativistic particles on Galactic distances) suggests that Galactic CRs are well coupled to the interstellar medium (ISM) and are likely to undergo diffusive motion, due to scattering off magnetic turbulence in the ISM. Most of current models of CR propagation treat the CR transport properties (such as the diffusion coefficient and the size of the propagation region) as fitting parameters and do not take into account the possible active role of CRs in determining them. In fact, the escape of CRs from the Galaxy leads to a gradient in the CR distribution function which can cause the excitation of Alfvén waves, due to CR streaming instability, that in turn determine the scattering properties of CRs, namely their diffusive transport. In addition, the CR pressure gradient can act as a force on the background plasma, directed away from the Galactic disk. If this force is large enough to win against gravity (due to Dark Matter, gas and stars), a wind can be launched, which affects the CR convective transport. The dynamics of CR-driven winds in the presence of self-generated turbulence is intrinsically non-linear. In fact, the CR density gradient determines the wind properties (velocity, pressure, magnetic field) and the excitation of the plasma waves which cause CR diffusion. On the other hand, Galactic winds could have a sizable effect on the CR distribution function, by advecting CRs out of the Galaxy, by influencing their spectral features, by affecting their radial distribution in the Galactic disk but also, possibly, by reaccelerating them at the wind termination shock (see Zirakashvili & Voelk (2006)). In addition, in such a scenario the CR diffusion coefficient, convection velocity and the size of the propagation region are not pre assigned, but rather they are derived self-consistently with the CR distribution function. The importance of Galactic winds is not only restricted to CR physics. In fact, galactic outflows have been observed in many galaxies and constitute an important ingredient in the galactic evolution. For instance, galactic winds affect the amount of gas available and pollute the galactic halos with hot plasma and metals, thus influencing the properties of the ISM and the star formation rate. As for the Milky Way, observations have not yet provided a clear answer as to the existence of such outflows, although the detection of absorption lines in the X-ray band (Oxygen OV II and OV III lines) show the presence of a hot dilute gas in the Galactic halo and the recent observation of the so called Fermi Bubbles in the Galactic Center region are likely to be connected with Galactic winds. Galactic winds may be powered by several mechanisms, for instance by thermal and radiation pressure gradients. However those mechanisms are unlikely to occur in the Milky Way, with the only possible exception of the Galactic Center region, since thermal and radiation pressure gradients are expected to be too small. Nevertheless, the CR pressure gradient may provide the force necessary to launch winds, making CR-driving an appealing mechanism for wind formation in our Galaxy. In this thesis we solve the coupled system of the hydrodynamic equations for CR-driven winds and for the CR transport in such winds. In our approach the CR transport is due to diffusion on self-generated Alfvén waves and to advection with these waves and with the wind. We then apply our solution method to the Milky Way and we investigate: 1) how the wind launching depends on the properties of the ISM (gas density and temperature, Galactic magnetic field), on the CR pressure and on the Galactic gravitational potential (including the Dark Matter halo); 2) the implications of CR-driven winds on the observed CR proton spectrum; 3) the effect of non-linear CR propagation in the presence of self-generated diffusion, both with and without CR-driven winds, on the CR distribution function in the Galaxy as a function of the Galactocentric distance, and we compare our predictions with the observed radial CR density and spectral slope, as inferred from observations of γ-rays.