ORIGIN OF COSMIC RAYS PASQUALE BLASI INAFOsservatorio Astrofisico

ORIGIN OF COSMIC RAYS PASQUALE BLASI INAF/Osservatorio Astrofisico di Arcetri GAMMA 400 - ICTP Trieste, May 2 -4 2013

OUTLINE q BASICS OF PARTICLE ACCELERATION AT SNR SHOCKS q GAMMA RAY EMISSION FROM ISOLATED SNR q GAMMA RAY EMISSION FROM CLOUDS q PROTON AND HELIUM SPECTRA AND COSMIC RAY PROPAGATION

BASICS OF PARTICLE ACCELERATION IN SNR PARTICLES GAIN ENERGY THROUGH REPEATED PASSAGES ACROSS THE SHOCK SURFACE THE DYNAMICAL REACTION OF ACCELERATED PARTICLES ON THE SHOCK CHANGE THE SHOCK STRUCTURE THE ACCELERATION PROCESS TURNS NON-LINEAR THE TRANSPORT OF ACCELERATED PARTICLES WITH THE SHOCK CREATES A CURRENT WHICH INDUCES CR-INDUCED PLASMA INSTABILITIES MAGNETIC FIELD AMPLIFICATION THE AMPLIFIED B-FIELD EXERTS A NON-LINEAR DYNAMICAL REACTION ON THE SHOCK ACCELERATED PARTICLES CHANGE THE CONDITIONS IN WHICH THEY GET ENERGIZED

COLLISIONLESS SHOCKS THE TIME SCALE FOR TWO PARTICLES COLLISIONS IS TOO LONG TO EXPLAIN THE FORMATION OF THESE SHOCK FRONTS SNR SHOCKS ARE ELECTROMAGNETIC (COLLECTIVE EFFECTS) MEDIATED BY INSTABILITIES IN THIS SENSE THEY ARE COLLISIONLESS AND MAY BE FELT ONLY BY ELECTRICALLY CHARGED PARTICLES

NON LINEAR THEORY: BASIC PREDICTIONS VELOCITY PROFILE COMPRESSION FACTOR BECOMES FUNCTION OF ENERGY 0 1 2 SPECTRA ARE NOT PERFECT POWER LAWS (CONCAVE) GAS BEHIND THE SHOCK IS COOLER FOR EFFICIENT SHOCK ACCELERATION EFFICIENT GROWTH OF B-FIELD IF ACCELERATION EFFICIENT SHOCK SELF-REGULATES !!!

X-RAY RIMS AND B-FIELD AMPLIFICATION TYPICAL THICKNESS OF FILAMENTS: ~ 10 -2 pc The synchrotron limited thickness is: In some cases the strong fields are confirmed by time variability of X-rays Uchiyama & Aharonian, 2007

Morlino & Caprioli 2012

MAXIMUM ENERGY Using the diffusion coefficient in the ISM derived from the B/C ratio: and the velocity of a SNR shock as u=5000 km/s one sees that: Too long for any useful acceleration NEED FOR ADDITIONAL TURBULENCE

PROTONS (thermal background) ELECTRONS (thermal background) SH K C O COSMIC RAY CURRENT IN THE PLASMA FRAME (positive charges) IN THE SHOCK FRAME CHARGE CONSERVATION NULL CURRENT

Linear Perturbations RESONANCE FOR LOW FREQUENCY WAVES ONE CAN HAVE A RESONANCE BETWEEN PARTICLES AND WAVES, IF POLARIZATION ALLOWS IT, WHEN:

RESONANT VS NON-RESONANT DIFFERENT UNSTABLE MODES EXIST SOME OF THEM CAN BE VERY FASTLY GROWING BUT ON SCALES MUSH LARGER ON SMALLER THAN LARMOR RADIUS THESE MODES LEAD TO LARGE B-FIELDS BUT SMALL SCATTERING RATE ONLY RESONANT MODES ARE EFFECTIVE IN SCATTERING THE PARTICLES THEREBY PROVIDING THE DIFFUSION NECESSARY TO INCREASE THE MAXIMUM ENERGY MANY OPEN QUESTIONS HERE --- THESE ARE THE VERY REASONS WHY WE ARE STILL DEBATING WHETHER SNR CAN BE THE SOURCES OF GALACTIC COSMIC RAYSS

GAMMA RAY EMISSION FROM SUPERNOVA REMNANTS

Tycho Supernova Remnant – 1572 SN Type Ia Distance ~3 kpc

THE CASE OF TYCHO Morlino&Caprioli 2011 STEEP SPECTRUM HARD TO EXPLAIN WITH LEPTONS DUE HERE TO THE ROLE OF THE SCATTERING CENTERS Emaxprotons ~ 500 Te. V

THE CASE OF TYCHO Morlino&Caprioli 2011

SOME POINTS… ü THE PION PEAK HAS SO FAR BEEN SEEN ONLY IN MOLECULAR CLOUDS ü THE DISCRIMINATION BETWEEN ICS AND p 0 DECAY CAN BE ACHIEVED BASED ON THE SPECTRUM ONLY FOR HIGH ANGULAR RESOLUTION o Different parts of the SNR may have different spectra reflecting a different origin or/and the presence/absence of a molecular cloud target (THIS MIGHT BE THE CASE FOR INSTANCE IN RXJ 1713) ü EXTENSION TO HIGH ENERGIES CAN PROVIDE A CHANCE TO SEE A CUTOFF IN THE PEV, BUT DO NOT BE TOO OPTIMISTIC (LOW PROBABILITY) ü IN ANY CASE, AT HIGH ENERGY, HARD TO COMPETE WITH CTA

THE CASE OF MOLECULAR CLOUDS Scenario n. 1: shock enters the cloud SNR Shock THE SHOCK BECOMES COLLISIONAL ON SCALES: THIS IMPLIES THAT THE SHOCK SLOWS DOWN SINCE IT FEELS THE LOAD OF THE GAS IN THE CLOUD (CONFIRMED BY OH MASER OBSERVATIONS) THE LARGE DENSITY OF NEUTRALS (LARGE TARGET DENSITY FOR pp) CAUSES EFFICIENT ION-NEUTRAL DAMPING WAVES ARE HARD TO EXCITE INSIDE THE CLOUD, THEREFORE ACCELERATION IS LIKELY TO BE SUPPRESSED AND PARTICLE SHOULD STREAM AWAY EASILY GAMMA RAY EMISSION FROM PARTICLES ACCELERATED AT PREVIOUS TIMES AND TRYING TO LEAVE

THE CASE OF MOLECULAR CLOUDS Scenario n. 2: shock away from the cloud SNR Shock THE SPECTRUM OF PARTICLES THAT REACHES THE CLOUD IS A FUNCTION OF THE AGE OF THE SNR AND AT ANY GIVEN TIME IT HAS A LOW ENERGY CUTOFF AT THE MAXIMUM ENERGY REACHED AT THAT GIVEN TIME n(p) TIME p

n(E) TIME E Behaviour of the cross section for pion production ~1/E (in the Feynman scaling regime) A LOW ENERGY CUTOFF IS IMPOSED BY FINITE TIME OF PROPAGATION GAMMA RAYS PRODUCED BY CR WITH E>Emin HAVE THE SAME SPECTRUM AS THE PRIMARY CR GAMMA RAYS PRODUCED BY CR WITH E<Emin HAVE A SPECTRUM REFLECTING THE BEHAVIOUR OF THE CROSS SECTION:

CLOUDS AS CALORIMETERS OF CR ESCAPE IS THE WEAK LINK BETWEEN ACCELERATION AND CR OBSERVED AT THE EARTH AT EACH TIME ONLY PARTICLES WITH E=EMAX(T) ESCAPE SPECTRUM IS FORMED BY TIME CONVOLUTION OF PEAKED SPECTRA GAMMA RAY SPECTRUM FROM NEARBY CLOUDS MAY PROBE THIS PHENOMENON BUT WHICH DIFFUSION? ? ? DIFFUSION LIKELY SELF-GENERATED AND NOT REPRESENTATIVE OF DIFFUSION IN THE ISM GAMMA RAY EMISSION ALSO SENSITIVE TO THE TOPOLOGY OF DIFFUSION: Giacinti et al. 2012, Nava & Gabici 2013 1 -D DIFFUSION 3 -D DIFFUSION

Nava & Gabici 2013

SPECTRA OF NUCLEI AND COSMIC RAY PROPAGATION 1. SPECTRA OF NUCLEI APPEAR TO HAVE A BREAK AT 200 GV 2. He NUCLEI SEEM TO HAVE A HARDER SPECTRUM THAN PROTONS

PAMELA CREAM

NON SEPARABLE D(E, z) THE STANDARD PREDICTION OF LEAKY-BOX-LIKE MODELS OF CR PROPAGATION THAT THE SPECTRUM SCALES AS n(E)~Q(E)/D(E)~E-g-d ONLY APPLIES WHEN THE DIFFUSION COEFFICIENT IS CONSTANT IN SPACE OR WHEN D(E, Z)=F(E)G(z). THE SIMPLEST EXAMPLE OF A NON-SEPARABLE D(E, z) IS THAT IN WHICH: D 2 H 2 D 1 H 1 THE TRANSPORT EQUATION IN ITS SIMPLEST VERSION IS:

THE INJECTION IS ASSUMED TO OCCUR ONLY IN AN INFINITELY THIN DISC: D(E) D 2 D 1 E IT IS EASY TO SOLVE THE TRANSPORT EQUATION IN THIS SIMPLE CASE. THE CR DISTRIBUTION FUNCTION IN THE DISC VICINITY IS: WHERE: EVEN IF THE DIFFUSION COEFFICIENT WITH DIFFERENT ENERGY DEPENDENCE IS HIGH IN THE HALO, THE SPECTRUM IN THE DISC SUFFERS A CHANGE OF SLOPE IN THE DIRECTION OF BECOMING FLATTER AT HIGH ENERGY (e. g. Dogiel 2001, Tomassetti 2012) ANISOTROPY FOLLOWS THE SAME BEHAVIOUR:

NON LINEAR EFFECTS INDUCED BY COSMIC RAYS COSMIC RAY GRADIENTS LEAD TO THE EXCITATION OF STREAMING INSTABILITY, WHENEVER THE STREAMING VELOCITY BECOMES >VA. THE GROWTH RATE OF THE INSTABILITY IS: FOR A SIMPLE INSTANCE OF TRANSPORT EQUATION (SEE ABOVE), THE SOLUTION IS: USING QLT THE DIFFUSION COEFFICIENT PARALLEL TO THE MAGNETIC FIELD IS: AND USING THESE EXPRESSIONS IN THE GROWTH RATE:

DAMPING vs CR-INDUCED GROWTH NL DAMPING AT ENERGIES BELOW 200 Ge. V THE WAVES GROW FASTER THAN THEY ARE DAMPED AND THE QUESTION OF THE SELF-GENERATED D(E) IS WELL POSED. NOTICE THAT TO THIS POINT WE USED NO MODEL, JUST OBSERVED QUANTITIES (FLUXES, GRAMMAGE, SLOPES)

TRANSPORT IN THE PRESENCE OF GROWTH AND DAMPING OF WAVES Power law in p a DENOTES THE TYPE OF NUCLEUS (BOTH PRIMARIES AND SECONDARIES ARE INCLUDED) THE EQUATION FOR THE WAVES IS: Damping as diffusion in k-space SUM OVER ALL NUCLEI OF THE RIGHT MOMENTUM TO GENERATE A WAVE WITH GIVEN k

PROTONS AND HELIUM NUCLEI Aloisio & PB 2013

SPECTRA OF HEAVIER NUCLEI Aloisio & PB 2013

GRAMMAGE AND DIFFUSION Aloisio & PB 2013

SECONDARY NUCLEI Aloisio & PB 2013

GAMMA RAYS FROM CLOUDS IN THE GOULD’S BELT (Neronov et al. 2012, Kachelriess & Ostapchenko 2012) Aloisio & PB 2013

SOME CONSIDERATIONS UNDERSTANDING OF THE PHYSICS OF THE KNEE (SPECTRUM AND CHEMICAL COMPOSITION) MAY BE INTRUMENTAL TO UNDERSTANDING THE NATURE OF THE TRANSITION FROM GALACTIC TO EXTRAGALACTIC COSMIC RAYS KASCADE-GRANDE, 2013

SUMMARY (1) u. A NEW GAMMA RAY INSTRUMENT FROM SPACE WOULD MAKE SENSE IF IT COULD AIM AT SEEING SNR BELOW 100 -1000 Ge. V WITH HIGH ANGULAR RESOLUTION THIS WOULD ALLOW TO HAVE SPATIALLY RESOLVED SPECTRA AND IDENTIFY THE REGIONS WHERE THE EMISSION IS REALLY DUE TO PION DECAYS (COINCIDENT WITH THE ONES WHERE THERE ARE FILAMENTS IN X-RAYS? ? ? ) u MORE OBSERVATIONS OF GAMMA RAYS FROM CLOUDS ILLUMINATED FROM SNR MAY SHED SOME LIGHT ON CR PROPAGATION AROUND SNR u IN THE 100 -1000 Te. V RANGE THE ISSUE OF COMPETITION WITH CTA SHOULD BE ADDRESSED u

SUMMARY (2) IF TO THINK OF AN INSTRUMENT FOR DETECTION OF NUCLEI… THE NEED IS TO CARRY OUT DIRECT MEASUREMENTS OF SPECTRA TO AS HIGH ENERGY AS POSSIBLE u BUT IT IS ALSO VERY IMPORTANT TO ASSESS THE EXISTENCE OF THE HARDENING IN ALL NUCLEAR SPECIES AT ROUGHLY THE SAME RIGIDITY (200 GV) – DETECTION OF THE WHOLE RANGE WITH THE SAME INSTRUMENT u GAMMA 400 WILL CERTAINLY HAVE SMALLER RESIDUAL GRAMMAGE DUE TO THE SPECIAL ORBIT u u THIS FEATURE SHOULD HOWEVER ALSO REFLECT IN ü ü ü THE SPECTRA OF GAMMA RAYS FROM INDIVIDUAL CLOUDS THE B/C RATIO AS A FUNCTION OF E/n (AMS? ) THE SPECTRUM OF ANTIPROTONS (AMS? )

BACKUP SLIDES

Resonant Instability Zweibel 1978, Bell 1978, Achterberg 1983, PB & Amato 2009 RESONANT MODES COSMIC RAY INDUCED THE UNSTABLE WAVES ARE ALFVEN WAVES WHOSE AMPLITUDE GROWS WITH A RATE PROPORTIONAL TO THE DENSITY OF RESONANT CR STRONG CR MODIFICATION THE UNSTABLE WAVES PROPAGATE FASTER THAN ALFVEN WAVES. THE GROWTH IS PROPORTIONAL TO NCR 1/2

Non-resonant high-k modes Bell 2004, Amato & PB 2009, Bykov 2011 Bell 2004; Amato & PB 2009

Non-resonant small k: firehose THIS INSTABILITY OPERATES WHEN YOU HAVE AN ANISOTROPIC PRESSURE CONTRIBUTION (ANISOTROPY AT THE SECOND ORDER IS REQUIRED IN THE PLASMA FRAME) IT FEELS THE WHOLE PRESSURE IN THE FORM OF CR IT AFFECTS SCALES MUCH LARGER THAN THE GYRORADIUS OF THE PARTICLES WITH THE HIGHEST ENERGIES

Fluid Instabilities Upstream Eddies in the upstream are twisted while crossing the shock surface (Richtmyer-Meshkov instability) This phenomenon leads to B-field amplification only downstream It does not reduce the acceleration time unless (perhaps) when the shock is quasi-perpendicular Giacalone & Jokipii 2007 THE MAIN DIFFERENCE BETWEEN THIS PROCESS AND THE CR INDUCED INSTABILITIES IS IN THE PRESENCE OF MAGNETIC FIELD AMPLIFICATION UPSTREAM --- ONLY CR ARE ABLE TO REACH UPSTREAM