Supernova Remnants HE Francesco Giordano University and INFN

Supernova Remnants @ HE Francesco Giordano University and INFN Bari Gamma 400 Workshop

Outline • Why @ HE (Emission Processes) • How Many @ HE (Towards the 1 st Catalog) • How Good (Study of systematic errors) • Different Families? • Conclusions 3/5/2013 F. Giordano @ Gamma 400 2

A “Typical” SNR SED n. H=100 cm-3 Kep=10 -1 Eb=10 Te. V 1=2 2=4. 5 B=100 u. G Fermi Range 0 Br e S m c yn on C I B CM n. H=1 cm-3 Scaling the environment density we pass from a p-p dominated SED to a IC dominated 3/5/2013 F. Giordano @ Gamma 400 3

Populations 3/5/2013 F. Giordano @ Gamma 400 4

Synchrotron radiation Zirakashvili and Aharonian (2007) A&A 465 ωc = 1. 5 Bp 2/(mc)3 3/5/2013 F. Giordano @ Gamma 400 5

IC Scattering Stuner et al. (1997) Ap. J 490 • U = 0. 26 e. V/cm 3 • T = 2. 73 K 3/5/2013 F. Giordano @ Gamma 400 6

Cosmic Microwave Background We used the following expression for the number density of CMB photon as a function of energy (blackbody photon distribution): where: • k is the Boltzmann constant and T = 2. 73 K; • The parameter U represents the energy density of CMB photons and it is equal to 0. 26 e. V cm-3. If we integrate over all values of energy we obtain the value of the number density: 3/5/2013 F. Giordano @ Gamma 400 7

Bremsstrahlung yield Baring et al. (2000) Ap. J 528 Density of environment Bethe-Heitler ultra-relativistic Cross Section 3/5/2013 F. Giordano @ Gamma 400 8

p-p interactions via 0 decay Parameterization proposed by Kamae et al 2006 3/5/2013 F. Giordano @ Gamma 400 9

Fermi-Detected SNRs 1 3 i d e n t i f i e d S N R Tycho 3/5/2013 Cygnus Loop CTB 37 A F. Giordano @ Gamma 400 counts Vela Jr 10

Detection of the p 0 -decay bump in SNRs IC 443 and W 44 are the two brightest SNRs in the Fermi-LAT range • The low energy break is very significant M. Ackermann et al. 2013 (~19σ and ~21σ for 60 Me. V E 2 Ge. V); • This gives unambiguous and robust detection of the pion decay bump • and clear proof that these SNRs accelerate protons. 3/5/2013 F. Giordano @ Gamma 400 11

Fermi-Detected SNRs + 1 3 4 3 i d e n t i f i e d S N R 2 F G L 3/5/2013 c a n d i d F. Giordano @ Gamma 400 aounts c 12

SNR Catalog: To better understand SNRs in a statistically significant manner within a MW context we: ● ● ● characterize the spatial and spectral morphology of all regions containing known SNRs. examine multi-wavelength (MW) correlation, including spectrum + morphology for radio, X-ray, and Te. V and CO, maser, IR, … determine statistically significant SNR classification(s) and perform spectral modeling 3/5/2013 F. Giordano @ Gamma 400 13

Characterize Ge. V Emission: Analysis Procedure Data Set: 3 years of P 7 SOURCE_V 6 LAT data • E: 1 -100 Ge. V • Region Of Interest: 10° around each SNR • Green’s Catalog (2009): • 274+5 SNRs Starting Model: • 2 FGL Overlapping sources? = None: Add a new extended source = 1 source (not PSR): Replace w extended source • > 1 source: Replace (non-PSR) source closest to radio centroid w extended source. Delete all other (non-PSR) sources. • • Localize source, fit extension • • • Disk extension seed = radio size Spectral model: power law Normalization of Galactic diffuse and all sources w/in 5 o of candidate are free during minimization procedure. Identifying SNRs via extension: If a SNR’s spatial extension is larger than Fermi’s PSF, we can detect its size, helping to positively identify it. 3/5/2013 F. Giordano @ Gamma 400 Output: • • Position, extension, significance Spectral energy distribution Region and residual maps Diagnostics 14

SNR Catalog: ● Fermi-LAT has the ability to spatially resolve a large number of the 279 known SNRs assuming their Ge. V and radio sizes are similar. 3/5/2013 F. Giordano @ Gamma 400 15

SNR Catalog: ● ● Fermi-LAT has the ability to spatially resolve a large number of the 279 known SNRs assuming their Ge. V and radio sizes are similar. Spatial extension measured for 15 SNRs, 6 of which are new candidates, permitting clear identification. 3/5/2013 F. Giordano @ Gamma 400 16

Systematic Error Study To explore systematic uncertainties related to the choice of the Interstellar Emission Model (IEM), we localized and fit 8 representative candidate SNRs using alternative IEMs. The eight remnants are a combination of hard and soft and point-like (x) and extended (o) sources and they are located in regions with different intensities of the IEM. 3/5/2013 F. Giordano @ Gamma 400 17

Alternative IEMs They are build using GALPROP with input parameters set as: • HI spin temperature =[150 K and optically thin], • CR source distribution =[SNR and Lorimer], • Halo height = [4 kpc and 10 kpc], and then fit to the data. The HI and CO emission split into 4 Galactocentric rings and the inverse Compton emission are fit simultaneously with the source of interest. Warning: • these 8 models do not span the complete uncertainty of the systematics. • the method for creating this model differs from that used to create the official Fermi-LAT interstellar emission model, so these 8 models do not bracket the official model. 3/5/2013 F. Giordano @ Gamma 400 18

Systematic Error Study SNR candidates' flux and index averaged over the alternative IEMs' solutions, compared to the standard (STD) model result. Flux: Index: P P R R SNRs: E E L L Our automated analysis finds a softer index and a much larger flux for SNR 347. 3 -0. 5 (RX J 1713) than that obtained in a dedicated analysis. I [Abdo et al. 2010] Since the best fit radius (0. 8 o I ) is larger than the dedicated o analysis’ (0. 55 ), the disk encompasses This make it softer than the M nearby sources that are not in the model. M more accurate analysis. 3/5/2013 F. Giordano @ Gamma 400 19 I I

E n d Radio-Ge. V Correlation? o f R a d i o s l i d e s h o w ● s y n c h r o t r o n 3/5/2013 ● Interacting SNRs: general correlation suggests a physical link Young SNRs show more scatter upper limits (i=2. 5, 99%) F. Giordano @ Gamma 400 20

Radio-Ge. V Index If radio and Ge. V emission arise from the same particle population(s), under simple assumptions, the Ge. V and radio indices should be correlated: ● ● Young SNRs: seem consistent Others, including interacting SNRs: softer than expected §p 0 =2 a+1 =a+1 d now challenge model Data e assumptions! c ● Underlying particle a populations may have y ● ● 3/5/2013 F. Giordano @ Gamma 400 different indices. Emitting particle o populations may not follow r a power law: breaks? Multiple emission zones? e + / - 21

Ge. V-Te. V Index IC 443 SED IC 443 IACT E 2 d. N/d. E Fermi Break region 1 Ge. V Indication of break at Te. V energies • Caveat: Te. V sources are not uniformly surveyed. F. Giordano @ Gamma 400 E 1 Te. V • 3/5/2013 22

Ge. V-Te. V Index RX J 1713 -3946 SED IACT E 2 d. N/d. E Fermi Break region RX J 1713 -3946 1 Ge. V Indication of break at Te. V energies • Caveat: Te. V sources are not uniformly surveyed. F. Giordano @ Gamma 400 E 1 Te. V • 3/5/2013 23

Environment? I n Young SNRs: t • Low Lg eevolving into low density r medium? a c Interacting SNRs: t • Higher Lg i encountering higher n densities? g 3/5/2013 S N R s F. Giordano @ Gamma 400 24

Or Evolution? Y o Due to u n • decreasing shock g allowing speed greater particle escape? S • decreasing maximum N acceleration energy R as SNRs age? s t e n d 3/5/2013 F. Giordano @ Gamma 400 25

May a simple simulation help? ? ? Particle injection =4. 5 3/5/2013 F. Giordano @ Gamma 400 26

Particle injection =4. 5 3/5/2013 F. Giordano @ Gamma 400 27

0 -> SEDs Particle injection =4. 5 1 3/5/2013 F. Giordano @ Gamma 400 28

Conclusions • Our systematic study has identified a statistically significant population of Galactic SNRs, including: – 6 new extended and >25 pointlike SNR candidates, – evidence for at least 2 SNRs’ classes: young and interacting. • Combining Ge. V and MW observations suggests that: – some SNRs' emitting particle populations may be linked, – simple model assumptions are no longer sufficient, allowing more complex models to be tested. • Improved observations and modeling will give us greater insight into SNRs, their acceleration mechanisms and their accelerated particles. • Accurately estimating SNRs’ aggregate particle acceleration mechanisms will also allow us to better quantify SNRs’ capability to produce the observed CRs. 3/5/2013 F. Giordano @ Gamma 400 29