Acceleration and Transport of Galactic Cosmic Rays Vladimir Ptuskin IZMIRAN, Moscow, Russia AMS days at CERN, April 15-17, 2015
CR spectrum at Earth GRB AGN interacting galaxies ultrafast pulsar E -2.7 cosmic ray halo, Galactic wind cosmological shocks Fermi bubble JxE 2 10 9 ev 10 20 ev LHC extragalactic 1/km 2 /century Sun pulsar, PWN SNR WMAP haze GC stellar wind close binary Galactic disk N cr = 10-10 cm -3 - total number density in the Galaxy w cr = 1.5 ev/cm 3 - energy density E max = 3x10 20 ev - max. detected energy Q cr = 10 41 erg/s power of Galactic CR sources A 1 ~ 10-3 dipole anisotropy at 1 100 TeV r g ~ 1 E/(Z 3 10 15 ev) pc - Larmor radius at B=3x10-6 G
Voyager 1 at the edge of interstellar space launched in 1977, 70 kb, 22 w 10 3 direct measurements of interstellar CR spectra at low energies 10 2 after E. Stone 2013 particle /(m 2 sec sr MeV/nuc) 10 1 10 0 10-1 10-2 10-3 10-4 He low energies: Voyager 1 Stone et al. 2013 high energies: BESS Pamela Sparvoli et al 2012 H 10-5 10 0 10 1 10 2 10 3 10 4 10 5 energy, MeV/nuc
cosmic ray intensity E -2.7 E -2.2 E -0.5 J cr (E) = Q cr (E) T e (E) source power escape time of CR from the Galaxy; ~ 10 8 yr at 1 GeV M51 ~ 15% of SN kinetic energy transfer to cosmic rays Ginzburg & Syrovatsky 1964 ν sn =(30 yr) -1 after Moskalenko traversed matter thickness (escape length) X e = m<n ism >T e ~ 12 g/cm 2 at 1 GeV/nuc (surface gas density of galactic disk ~ 2.5 10-3 g/cm 2 )
NGC4631 basic empirical diffusion model Ginzburg & Ptuskin 1976, Berezinskii et al. 1990, Strong & Moskalenko 1998 (GALPROP), Donato et al 2002, Ptuskin et al. 2006, Strong et al. 2007, Vladimirov et al. 2010, Bernardo et al. 2010, Maurin et al 2010, Putze et al 2010, Trotta et al 2011 H = 4 kpc, R = 20 kpc empirical diffusion coefficient D 28 2 ~ 310 cm / s at 1 GeV/n diffusion mean free path escape length l 1 pc X vh 2D e
diffusion-convection transport equation for cosmic rays Ginzburg & Syrovatsky 1964, Skilling 1975, Berezinskii et al. 1990 (loaded in GALPROP code) fi u f i 1 2 fi Difi ufi p p K 2 i t 3 p p p p 1 spatial diffusion advection stochastic acceleration dp f v. 2 i 2 p fin i fi qi Sij p p dt iion, i ji f( t, r, p) - phase-space density (isotropic part); N t dpp f ionization energy loss fragmentation radioactive decay 2 cr (, r) 4 - total cosmic-ray number density; N t E p f p 2 cr (, r, kin) 4 ( ) / v - spectrum on kinetic energy; 2 Jt (, r, E) 4 pf( p) - cosmic ray intensity. source contribution of fragmentations and decays
r g = 1/k - resonance condition spectrum of turbulence mechanism of cosmic ray diffusion 2 res D vr 3 g B = w(k)dk k res B B 2 tot 2 res Jokipii 1966, cosmic ray particles are strongly magnetized but large-scale random magnetic field (L max ~ 100 pc) makes diffusion close to isotropic
interstellar turbulence Armstrong et al 1995 10 17 ev 10 9 ev Kolmogorov-type spectrum L max 100 pc, B 3 G, A ( B/ B) 1, L L 27 1/3 p 2 17 D 410 cm /s, E 10 Z ev Z GV Kraichnan type spectrum A L p 0.3, D Z 1/2 w(k) ~ k -5/3 k -3/2 Kolmogorov Kraichnan Lin et al. 2015
flat component of secondary nuclei produced by strong SNR shocks Wandel et al. 1987, Berezhko et al. 2003 production by primaries inside SNRs grammage gained in SNR N N 2,flat 2,stand ~ X X SNR ISM ~ 0.02 reacceleration in ISM by strong shocks N N 2,flat 2,stand X ~ ISM fsnr XSNR ~ 0.2 grammage gained in interstellar gas Berezhko et al. 2003 RUNJOB 2003 preliminary volume filling factor of SNRs standard plain diff. reacceleration plain diff. reacceleration n ism = 0.003 1 cm -3 Bohm diffusion T SNR = 10 5 yr
regimes of cosmic ray diffusion Kolmogorov spectrum: D ~ vr 1/3 + reacceleration - may be valid up to 10 17 Z ev -source spectrum R -2.4 at high energies too soft to be produced in SNRs? (dispersion of SNR parameters may help); considerable flattening at < 3 GV - anisotropy acceptable - peak of B/C ratio at 1 GeV/n explained by distributed reacceleration - underpredicts antiproton flux below few GeV - in contradiction with theory of MHD turbulence (Alfven wave distribution in k-space) Kraichnan spectrum: D ~ vr 1/2 + wave damping - may be valid up to 10 16 Z ev only -source spectrum R -2.2 consistent with shock acceleration in SNRs - anisotropy too high - peak of B/C ratio at 1 GeV/n qualitatively explained by turbulence dissipation on CR - not in contradiction with theory of MHD turbulence (fast magnetosonic waves)
diffusion-convection model: galactic wind driven by CR H ef (p/z) u inf = 500km/s R sh = 300 kpc Ipavich 1975, Breitschwerdt et al. 1991, 1993, Everett et al.2008, 2010, Salem & Bryen 2014 CR scale height is larger then the scale height of thermal gas. CR pressure gradient drives the wind. + cosmic ray streaming instability with nonlinear saturation Zirakashvili et al. 1996, 2002, 2005, VP et al. 1997, 2000, D 1 1.1 s vb p p q cr Zmpc Zmpc 27 2 ~ 10 cm / s, (3 1) / 2 2.7, at 2.1 s s X s 1 2 H ef p p ~ ~ ~ D Zmpc Z 0.55
u sh SNR shock diffusive shock acceleration Fermi 1949, Krymsky 1977, Bell 1978, -γ s σ +2 J~p, γ s = = 2 σ -1 compression ratio = 4 for test particles! ushr sh >10 D(p) -condition of CR acceleration and confinement -D(р) should be anomalously small both upstream and downstream; CR streaming creates turbulence in shock precursor Bell 1978; Lagage & Cesarsky 1983; McKenzie & Vőlk 1982 Bohm limit D B =vr g /3: ush Emax 0.3Ze BR c E max,ism =10 13 10 14 Z ev ~ B sh t -1/5 at Sedov stage sh for B ism = 5 10-6 G
more detailed investigation: - streaming instability gives B >> B ism in young SNR Bell & Lucek 2000, Bell 2004, Pelletier et al 2006; Amato & Blasi 2006; Zirakashvili & VP 2008; Vladimirov et al 2009; Gargate & Spitkovsky 2011, Bell 2013, Caprioli & Spitkovsky 2014 from Caprioly 2012 confirmed by X-ray observations B 2 /8π = 0.035 ρu 2 /2 Voelk et al. 2005 under extreme conditions (e.g. SN1998 bw): E max ~ 10 17 ZeV, B max ~ 10-3 G -wave dissipation in shock precursor leads to rapid decrease of E max with time VP & VZ 2003 Jxp 2 - finite V a leads to steeper CR spectrum u u V 1 a,1 V 2 a,2 t = 10 3 yr Alfv. drift downstream
effect of cosmic ray pressure 2 P ξ ρu, ξ 0.5 cr cr sh cr - back reaction of cosmic-ray pressure modifies the shock and produces concave particle spectrum Axford 1977, 1981; Eichler 1984; Berezhko et al. 1996, Malkov et al. 2000; Blasi 2005 σ 2.5 sub approximate modified spectrum N (E ) E -2 cr kin kin - Eichler spectrum
equations for numerical modeling Zirakashvili & VP 2008-2012,
particle acceleration and radiation in SNR Berezhko et al. 1994-2006, Kang & Jones 2006 Zirakashvili & VP 2006-2012, semianalytic models Blasi et al.(2005), Ellison et al. (2010) ) Zirakashvili & VP 2012 Zirakashvili et al 2014 Cas A radio polarization in red (VLA), X-rays in green (CHANDRA), optical in blue (HST)
cosmic ray spectrum injected in ISM during SNR life time VP, Zirakashvili, Seo 2010 - Bohm diffusion in amplified magnetic field B 2 /8π = 0.035 ρu 2 /2 ( Voelk et al. 2005 empirical; Bell 2004, Zirakashvili & VP 2008 theoretical) JxE 2 - account for Alfvenic drift w = u + V a upstream and downstream - relative SNR rates: SN Ia : IIP : Ib/c : IIb = 0.32 : 0.44 : 0.22 : 0.02 Chevalier 2004, Leaman 2008, Smart et al 2009 protons only E -3 «knee» is formed at the beginning of Sedov stage E Z=1.1 10 W n M ev, 15 1/6-2/ 3 knee sn,51 ej 15-1 knee sn,51-5 w,6 ej E Z=8.4 10 W M u M ev
calculated spectrum of Galactic cosmic rays pc JxE 2.75 (normalized at 10 3 GeV; ISM diffusion D Ze ) VP, Zirakashvili, Seo 2010 0.54 solar modulation interstellar spectrum of all particles extragalactic component data from HEAO 3, AMS, BESS TeV, ATIC 2, TRACER experiments data from ATIC 1/2, Sokol, JACEE, Tibet, HEGRA, Tunka, KASCADE, HiRes and Auger experiments
features to explain: hardening above 200 GeV/nucleon superposition of sources Vladimirov et al 2012 new source Zatsepin & Sokolskaya 2006 reacceleration in local bubble Erlykin & Wolfendale 2011 streeming instability below 200 GV Blasi et al. 2012 spectra of p and He are different shock goes through material enriched in He: -helium wind VP et al. 2010, - bubble Ohira & Ioka 2011 effect of injection Malkov et al 2011 JxE 2.7 possible explanation of both features: concave spectrum and contribution of reversed SNR shock VP et al 2013
collection of data Mitchell 2013 positrons in cosmic rays AMS, 2014 electrons positrons
- pulasar/pwn origin Harding, Ramaty 1987, Aharonian et al. 1995, Hooper et al. 2008, Malyshev et al. 2009, Blasi, Amato 2011, Di Mauro et al. 2014 - e+ production and acceleration in shell SNRs Blasi 2009, Mertsch & Sarkar 2014 - reverse shock in radioactive ejecta Ellison et al 1990, Zirakashvili, Aharonian 2011 - annihilation and decay of dark matter Tylka 1989, Fan et al 2011 electrons positrons anti-matter factory electrons + positrons positron fraction L pairs = 20-30% L psr Di Mauro et al 2014: SNR (electrons), PWN (electrons + positrons), secondary leptons from ISM after Amato
structure above the knee JxE 3 p,he knee 2nd knee knee at 3 PeV (light primaries), hardening at 20 PeV (medium component), 2 nd knee at 300 PeV Prosin et al 2013 Sveshnikova et al 2014: composition at 1PeV: H 17%, He 46%, CNO 8%, Fe 16% + EG component H 75%, He 25% as in Kotera & Lemoine 2008
JxE 2.7 CNO,Si,Fe p,he Iron knee at 8*10 16 ev = 26*3*10 15 ev TUNKA KASCADE-G CNO,Si,Fe light ankle at 1.2*10 17 ev - transition to EG component or onset of a new high energy Galactic source population p,he
other Galactic accelerators at ultra-high energies reacceleration by multiple shocks SNR OB association u=3 10 3 km/s B=10-5 G R=30 pc SNR SNR R u f ~ 1/p 3 t a ~ R/(F sh u) at D i < ur ~ D/(F sh u 2 ) at D i > ur E max ~ 10 17 Z ev Bykov & Toptygin 1990, 2001, Klepach et al. 2000 newly-born pulsars with ultrarelativistic wind E ~ 10 19 Z B 13 (Ω/3000 rad/s) 2 ev Blasi et al. 2000, Arons 2003, Fang et al. 2013,
Fermi bubbles Su et al 2010 gamma-ray emission E ph = 1 100 GeV, 4x10 37 erg/s J e ~ E -2 gamma rays, X-rays, radio, WMAP Haze
Galactic wind u SNR R acceleration at termination shock Jokipii & Morfill 1985, 1991 Zirakashvili & Völk 2004 R = 300 kpc, u = 400 km/s E max = 10 18 Z ev galactic disk
Kulikov & Kristiansen 1958 GZK suppression? from Beatty et al 2013
Conclusions Cosmic ray origin scenario where supernova remnants serve as principle accelerators of cosmic rays in the Galaxy is strongly confirmed by numerical simulations. SNRs can provide cosmic ray acceleration up to 5x10 18 ev. PWN may be responsible for observed flux of cosmic ray positrons. Diffusion provides reasonably good description of cosmic ray propagation in the Galaxy even under simple assumptions on cosmic ray transport coefficients and geometry of propagation region (e.g. as loaded in GALPROP code).