1. Acceleration of GCR in stellar winds and the GCR 22 Ne/ 20 Ne source ratio N. Prantzos (Institut d’Astrophysique de Paris)

2. Galactic Cosmic Ray Source Composition CNO overabundant by ~1.5 to 8 ; Most excess CNO attributed to WR stars No, for elemental ratios  Selection effects Refractories: overabundant, but no clear trend with A/Q Ellison, Meyer, Drury (1997) : SN shocks accelerate ISM gas (volatiles) and sputtered grains (refractories) After taking into account several selection effects, it seems that the Source composition of GCR today is ~solar. Except for some excess C and Ne22/Ne20 from WR star winds ( Cassé and Paul 1982) Is it solar ? Yes, for most isotopic ratios Volatiles : elements with high A/Q (mass to charge) favored

3. Galactic Cosmic Rays : what is the composition of accelerated matter ? SBM SN ISM Wind C, N, Ne22.. SN ISM FS ISMB A D C FS RS SN Forward Shock accelerates ISMReverse Shock accelerates SN ejecta SN Shocks accelerate Superbubble matterForward Shock accelerates Wind + ISM

4. A forward shock is launched at M EXP and runs through the wind of the star, enriched with products of H- and/or He- burning, through the unprocessed envelope, and finally in the interstellar medium. ASSUMPTION: Particle acceleration starts in the beginning of the Sedov-Taylor (ST) phase, when M SWEPT ~ M EJEC (BUT: When does it stop ?) Depending on the previous mass loss of the star, acceleration may occur when the shock is still within the wind (more massive stars) or in the ISM (less massive stars), thus affecting the composition of accelerated particles.

5. Ingredients of stellar models (convection, mass loss, rotation etc.) determine the various quantities (M EXP, M REM, M WIND, etc.) Models with rotation have larger amounts of processed material M PROC MASS (M ⊙ ) Stellar models with mass loss and rotation ( Hirschi et al. 2005 ) or no rotation ( Chieffi and Limongi 2006 )

6. MASS (M ⊙ ) Rotating models have larger yields of He-burning products (C-12, Ne-22, etc.) in their winds

7. Integrated mass swept up by the forward shock, as it moves outwards Stellar models with mass loss and rotation (Hirschi et al. 2005)

8. (van Marle et al. 2008) Fast rotating Z=0.001 20 M ⊙ The circumstellar environment of mass losing stars For a steady stellar wind, the density profile can be approximated by Density  Radius -2 with

9. Propagation of forward shock into a stellar wind of profile ρ(r)  r -2 (Ptuskin and Zirakhasvili 2005, Caprioli 2011 )

10. Caprioli 2011

11. TIME (yr) Particle acceleration starts in beginning of ST and is assumed to stop when the velocity of the shock drops to  MIN chosen such as the IMF averaged ratio Ne22/Ne20 of accelerated particles equals the observed one R = (Ne22/Ne20) GCR = 5.3 ⊙  

12. MASS (M ⊙ ) The IMF averaged ratio Ne22/Ne20 of accelerated particles equals the observed one R = (Ne22/Ne20) GCR = 5.3 ⊙ for  MIN =1900 km/s for rotating models and 2400 km/s for non-rotating ones HMM05 CL06

13. A1: Beginning ST and acceleration A2: End acceleration  S =1900 km/s The forward shock accelerates particles from a pool of mass M ACC = A2 – A1 between the beginning of ST ( A1 ) and  =1900 km/s ( A2 ) Energy of accelerated particles Mass of accelerated particles Efficiency of acceleration Processed material

14. Are GCR accelerated in Superbubbles ? SN eject at explosion much larger amounts of Ne20 than in their winds but they eject amounts of Ne22 comparable to the ones of their winds The Ne22/Ne20 ratio produced by a stellar generation (averaged over the IMF) should be close to solar (if not: problems with global Nucleosynthesis) In a superbubble, material from the whole IMF should be mixed and accelerated

15. In a superbubble the time integrated (GCR) Ne22/Ne20 ratio remains compatible with the observed GCR one ONLY for a short early period and ONLY if no original material is left over. Most of the time, and in realistic conditions Ne22/Ne20 is close to solar Superbubbles are NOT at the origin of GCR Ne22/Ne20 nor at the origin of GCR. ( Contrary to Higdon et al. 1998, Higdon and Lingenfelter 2003, Binns et al. 2005, 2008 )