1. Modeling photon and neutrino emission from the supernova remnant RX J1713.7-3946  Constraints from geometry  Constraints from spectral energy distribution  Ingredients for a physical model  Results and neutrino predictions Jean Ballet (SAp, CEA Saclay)Rencontres de Blois, Wednesday May 21 st, 2008 with Gilles Maurin (KM3NeT postdoc) and Gamil Cassam-Chenaï

2. Rationale SNRs are the most likely source of Galactic cosmic-rays on theoretical grounds (OB associations might be even better, but more diffuse) Good observational evidence (radio and X-ray synchrotron from electrons, TeV emission) Must be the place in the Galaxy where the density of TeV to PeV hadrons is largest Good target for neutrino astronomy, if there is enough gas around Let us look at the best known TeV SNR, RX J1713.7-3946

3. RX J1713.7-3946 1 degree diameter remnant close to Galactic plane (G347.3-0.5). Average absorbing column (from X-rays) 5 to 6 10 21 cm -2. Likely distance is 1 to 1.5 kpc (association with clouds in the West and absorption value). Radius is then 8 to 13 pc. Might be remnant of SN 393 (1600 years old). Central compact object is present, therefore SN II. Possibly exploded in wind-blown shell recently reached by the shock. No thermal emission detected. Most likely reason that the ambient density is low (< 0.02 cm -3 ). Consistent with the size for reasonable energy (10 51 erg). X-rays (excluding point sources) are synchrotron, due to electrons accelerated at TeV energies. Emission is filamentary (probably sheets in projection). If width (40” or 0.25 pc) is interpreted as cooling length, implies post-shock B around 80 μG.

4. HESS Point source XMM-Newton mosaic Acero et al 2008 Central Compact Object

5. Parameters  Supernova: age (t 0 = 1600 yrs), energy (E = 10 51 erg), ejected mass (10 M o )  Local conditions: density (n 0 ), distance (1 kpc)  Particle acceleration: injected fraction (  inj =5 10 -4 ), electron/proton (K ep ), magnetic field (B 0 ) following Berezhko and Ellison 1999, ApJ 526, 385 Constraints on global parameters Constraints  Angular size (E/n 0, t 0 )  Expansion over time or Doppler width: shock velocity (E/n 0, t 0 )  Thermal X-ray emission (n 0 )  Synchrotron emission level (B 0,  inj, K ep )  X-ray synchrotron rim width (B 0 )  Width between ejecta and blast wave (  inj, B 0 )

6. Modeling supernova remnants Analytic (1D self- similar) hydrodynamics Ionized hot gasShock accelerated particles Accelerated particles throughoutThermal spectrum Non-thermal spectrum3D (X,Y,E) model Acceleration Ionization, electron heating Propagation Cooling Emission Projection Applied to Tycho SNR (Cassam-Chenaï et al 2007, ApJ 665, 315)

7. Young SNRs: Hydrodynamics Power law density profiles => self-similar solutions. Can accommodate stellar winds and represent approximately shell encounter (ρ as r 5 for example) Initial conditions : Arnett 1988, ApJ 331, 377 Ejecta Chevalier 1983, ApJ 272, 765; Decourchelle et al 2000, ApJL 543, 57 ejecta Reverse shock ISM Forward shock

8. Leptonic model ASCA ATCA H.E.S.S. IC Synchrotron Pions E(eV) E 2 dN/dE (eV.cm -2.s -1 ) Flat ambient density n 0 = 8 10 -3 cm -3 Distance D = 1 kpc Electron/proton = 10 -2 Mag field B 0 = 3 μG M shocked = 0.6 M o E pmax = 40 TeV = 0.8 keV (ejecta) Parameters OK except magnetic field (X-ray filaments) B field could be larger if B turbulence decays behind shock (Pohl et al 2005, ApJ 626, L101) so that volume for synchrotron is smaller. Allowed by radio. Non thermal spectral fit not very good (spectrum too peaked) as in Aharonian et al. 2006, A&A 449, 223

9. Hadronic model ASCA ATCA H.E.S.S. IC Synchrotron Pions E(eV) E 2 dN/dE (eV.cm -2.s -1 ) Flat ambient density n 0 = 0.3 cm -3 Distance D = 1 kpc electron/proton = 8 10 -4 mag field B 0 = 12 μG M shocked = 7.3 M o E pmax = 70 TeV = 1.6 keV Remnant is too small at E = 10 51 erg (40’ diameter) Non thermal spectral fit rather good (fits slope OK) as in Berezhko & Völk 2006, A&A 451, 981 Predicted thermal emission way too high (as Katz and Waxman 2008, JCAP 1, 18) Shell model only marginally better Most of the gas must be outside SNR and cold as in Malkov et al. 2005 (ApJ 624, L37). Predicts harder spectrum (energy-dependent diffusion ahead of the shock, not in our code now).

10. GeV and neutrino emission H.E.S.S. GLAST 5 years. Leptonic Hadronic Gamma-rays GLAST would see hadronic source in 1 year (but diffusion into neighbouring clouds will not be so favourable) H.E.S.S.-2 will see whether spectrum is harder at 100 GeV than at 1 TeV E(eV) E 2 dN/dE (eV.cm -2.s -1 ) H.E.S.S.-2 North hemisphere Extended source for KM3NeT physics case (preliminary)

11. Modelling photon and neutrino emission from the supernova remnant RX J1713.7-3946  Adapted approximate (1D self-similar) but self-consistent SNR model to predict  –ray and neutrino emission  Computes accurately thermal X-ray emission  Applied to RX J1713.7-3946: leptonic model can work, hadronic model requires target gas to be cold (diffusion ahead of the shock)  Neutrino emission expected in hadronic model Jean Ballet (SAp, CEA Saclay)Rencontres de Blois, Wednesday May 21 st, 2008 with Gilles Maurin and Gamil Cassam-Chenaï