B. Degrange ICATPP, October 2010 1 Recent results from H.E.S.S. Bernard Degrange, for the H.E.S.S. collaboration LLR, Ecole polytechnique & CNRS/IN2P3.

B. Degrange ICATPP, October 2010 1

Recent results from H.E.S.S.

Bernard Degrange, for the H.E.S.S. collaboration LLR, Ecole polytechnique & CNRS/IN2P3


Since 2003, the H.E.S.S. array has been unveiling the southern TeV gamma-ray sky

Stereoscopic array of 4 atmospheric Cherenkov telescopes

Khomas Highlands (Namibia, 23°S, 15°E, 1800 m a.s.l.)

Large mirrors (13 m diameter dishes)

Fine grain cameras (960 pixels/camera), fast trigger electronics located in the camera, 5° field of view well adapted to extended sources

Energy threshold : from 120 GeV (zenith) to 700 GeV at 60° from zenith


H.E.S.S. sensitiviy at zenith : 1% of the flux of the Crab nebula is detected in 25 hours with a 5σ significance for a point-like source.

Angular resolution Δθ ≈ 0.07° , always < 0.1°

Gamma-ray sources can be located with a precision of 1 arc min. or better

Advanced reconstruction methods have improved sensitivity (higher background rejection) and angular resolution

The present H.E.S.S. catalogue includes > 60 sources of TeV gamma-rays

M. de Naurois & L. Rolland

Astropart. Phys. 2009


Galactic sources : new shell-type supernova remnants

Normal nearby galaxies : LMC and NGC 253

Active Galactic Nuclei : radio-galaxies and blazars

Towards HESS phase II and conclusion

Outline

A few recent H.E.S.S. results


Inner Galaxy (2004) + extension (2005-2008) from ℓ = -85° (or 275°) to ℓ = +60° and |b|<3°

Survey is complete for fluxes > 0.09 Crab (+ deeper observations)Low diffuse flux (≠ Fermi/AGILE) → Individual sources appear clearlyMost of the revealed sources are mildly extended (D > 3 ' to 4 ')

G.C.

1. Galactic sources: new shell-type supernova remnants

Sun


1.1 The remnant of a historical supernova : SN 1006

The explosion took place slightly above the Galactic Plane in a rather uniform medium and magnetic field, as attested by the morphology in two polar caps observed in X-rays (XMM-Newton, Rothenflug et al. 2004)

The morphology likely reflects the NE-SW orientation of the average magnetic field, suggesting that parallel shocks (i.e. B // shock normal) lead to efficient acceleration (Völk et al. 2003).

However, the ambient density of interstellar matter is low (≈ 0.05 cm-3) (Acero et al. 2007) and 130 hours of observations were necessary for H.E.S.S. to firmly detect the source A&A 516 (2010)A62


H.E.S.S. results on SN 1006 : morphologyA&A 516 (2010)A62

White contours: constant X-ray intensity derived from XMM-Newton flux map and smoothed to the

H.E.S.S. point-spread function, enclosing 80%,60%,40%,20% of the X-ray emission

Radial profile (NE-SW)

Azimuthal profile


Very low fluxes (< 1% Crab nebula) NE Φ(>1TeV) = (0.233 ± 0.043 ± 0.047) × 10-12 cm-2 s-1

SW Φ(>1TeV) = (0.155 ± 0.037 ± 0.031) × 10-12 cm-2 s-1

Power-law spectra from both lobes with similar indicesΓ = 2.35 ± 0.14 ± 0.2Γ = 2.29 ± 0.18 ± 0.2

H.E.S.S. results on SN 1006 : energy spectrum A&A 516 (2010)A62

E2 dΦ/dE dΦ/dE

Mixed leptonic/hadronic scenario

• Leptonic scenario: difficulty to reproduce the H.E.S.S spectrum.

• Hadronic scenario: too much energy in the protons to account for the H.E.S.S spectrum.

• Mixed scenario: satisfactory

But no scenario can be really excluded


1.2 A new shell-type supernova remnant: the formerly unidentified source HESS J1731-347

Discovered in the H.E.S.S. galactic plane survey as an unidentified source (Aharonian et al. 2008)Tian et al. 2008 find a faint shell-like structure in radio (ATCA & Parkes) and in X-rays (ROSAT) coincident with the H.E.S.S. source (G353.6-0.7) 0.4° diameter, age estimated to 27 kyrDeeper X-ray observations (Acero et al. ICRC 2009) with Suzaku, XMM and Chandra → interaction with a molecular cloud in foreground → distance > 3.2 kpcH.E.S.S. deeper observations (60 hr) detect a new shell (22 σ) in very-high-energy gamma-rays; flux ~ 9% Crab

Aharonian et al. 2008

Tian et al. 2008

F. Acero et al.,

COSPAR 2010


HESS J1731-347 : a 5th shell detected at very high energies, the first whose discovery was based on

gamma-ray observations

The radial profile clearly suggests a shell (filled sphere excluded at 3.9σ)Distance estimated to be > 3.2 kpc :

X-ray show a strong gradient (SE-NW) of absorption → molecular cloud in foreground12CO survey data (CfA) can explain the absorption column density obtained from X-rays by a distance of 3.2 kpc

At such a distance, the luminosity of the source (> 1.07 × 1034 erg s-1) is greater than that of RX J1713-3946Source B could be due to cosmic ray interactions in a nearby molecular cloud.

Acero et al. COSPAR 2010


2. Nearby normal galaxies The Large Magellanic Cloud

(D=50kpc) The starburst galaxy NGC 253

(2.6 Mpc<D<3.9 Mpc)


2.1 A young pulsar-wind nebula, N157B, in the Large

Magellanic Cloud (50 kpc)

47 hours of good-quality data revealed the young (≈5000 years) pulsar-wind nebula N157B (15 σ) at the border of the star-forming region 30 Doradus

The source is point-like and well separated from SN1987A.

This nebula, observed in radio and X-rays (Chandra), is powered by pulsar PSR J0537-6910, which has the highest spin-down luminosity 4.8 × 1038 erg s-1

Gamma-ray energy flux (1-10 TeV) = 1.4 × 10-12 erg cm-2 s-1. from N157B ~2% Crab

Upper limit on the energy flux from SN1987A (>1 TeV) : < 5 × 10-13 erg cm-2 s-1 (99%CL) is a factor of ~2 above the values predicted by Berezhko & Ksenofontov 2006.

N. Komin et al. COSPAR 2010


Bubbles created by many stellar explosions and stellar winds may coalesce, thus creating a super-bubble whose shock wave can accelerate particles over a longer period than that of an isolated supernova → particle acceleration up to > PeV ?

Superbubbles are difficult to observe in the Milky Way, due to their wide angular extension.

Solution : observe a nearby galaxy with a boosted formation rate of massive stars → increased rate of supernovae: starburst galaxy.

H.E.S.S. detected NGC 253: 5.2 σ in 119h of observation. Flux (>220 GeV) ≈ 0.3% that of the Crab nebula is concentrated in the nucleus

Density of protons (E>1.3 TeV) in the 100 central parsecs = 1400 × that in the central region of the Milky Way.

2.2 The starburst galaxy NGC 253

Science 326 (2009) 1080

2.6 Mpc<D<3.9 Mpc


To date, H.E.S.S. has detected 15 AGN All these sources are radio-loud AGN (~5% of all AGN) ; the radio emision is due to the relativistic jets ejected from the central region.All AGN from the H.E.S.S. catalogue but 2 belong to the “blazar” class, whose characteristics result from an observation effect: jets are emitted at small angle with respect to the line of sight.The 2 exceptions are radio-galaxies:

M87 (HEGRA, H.E.S.S., MAGIC, VERITAS) variability on day-scale confirmed → size of emission region <0.15 pc Flux ≈ 1% to 5% that of the Crab nebula → flare 10% Crab in 2008Centaurus A (H.E.S.S.) : Flux ≈ 0.8% that of the Crab nebula

3. Active Galactic Nuclei (AGN)

Courtesy of C.M. Urry and P. Padovani


Two radio-galaxies: emission region compatible with radio core

GalaxyDistance

(Mpc)

Angle (jet-l.o.s.)

Black hole mass (solar masses)

Photon index Detection

M87 16.7 30° (6.0±0.5) × 109 2.62±0.35 (2004) 2.22±0.15 (2005)

HEGRA,HESS, VERITAS,

MAGIC

Centaurus A 3.7 15°-80° (5.5±3) × 107 2.7±0.5±0.2 HESS

M87 Multi-wavelength campaign 2008 H.E.S.S./MAGIC//VERITAS + VLBA (43 GHz)

+Chandra: Science 325 (2009) 444 emission likely due to core

innermost resolved knot HST1 in jet excluded (low state of HST1 in X-rays during largest

γ-ray flare in Feb. 2008 (10% Crab))

Chandra

Cen A: 5 σ in 115 h of observation Flux (E>250GeV) ~ 0.8% Crab

Cen A: ApJ 695(2009)L40

E2 dΦ/dE


Blazars are characterized by a two-bump broad band spectrum: e.g. Optical (ATOM)-X-ray (Swift, RXTE)-Fermi LAT-H.E.S.S.

on the blazar PKS 2155-304 (z=0.117) in a low state in 2008

Blazars detected by Cherenkov arrays are mostly « High-frequency-peaked BL Lac objects» or HBLSimultaneous multi-wavelength campaign on PKS2155-304 in 2008 reveals a typical HBL spectrum:

synchrotron bump peaks in UV/X,γ-ray bump peaks in the 100 GeV range

August 25 to September 6, 2008

ApJ 696 (2009) L150

E2 dΦ/dE


GalaxyRedshift

zBlazar class

Flux (%Crab)

Photon index Γ

Variability

Mkn 421 0.030 HBL ~ 300 2.1±0.1 < hour

AP Librae 0.049 LBL ~ 2

PKS 0548-322 0.069 HBL 1.4 2.86±0.34

PKS 2005-389 0.071 HBL 2.8 4.0±0.4 ~ month

RGB J0152+017 0.080 HBL 2.0 2.95±0.36 ~ month

PKS 2155-304 0.116 HBL 15 to 1500 3.53±0.06 ~ 3 min.

1ES 0229+200 0.139 HBL 1.8 2.50±0.19

H 2356-309 0.165 HBL 2.3 3.09±0.24 ~ month

1ES 1101-232 0.186 HBL 2.3 2.94±0.20 ~ year

1ES 0347-121 0.188 HBL 2.0 3.10±0.23 ~ year

PKS 0447-439 0.205 HBL 1.0 4.47±0.37

1ES 0414+009 0.287 HBL 1.0 3.53±0.39

PG 1553+113 >0.25 HBL 3.4 4.5±0.3

Blazars detected

by H.E.S.S., as of 2010

Recent detections

are in boldface

characters

Blazars not

detected by Fermi-LAT are in purple


Blazar variability: the case of PKS 2155-304

PKS 2155-304: most of the time, flux ~ 15% Crab (“quiescent” regime)In July 2006, flaring regime during 4 nights with extreme (minute scale) variability ApJ 664 (2007) L71

1st night MJD53944

Light curve (monthly averages) 2005-2007arXiv:1005.3702


PKS 2155-304 variability (arXiv:1005.3702)

Flux distribution (2005-2007) excluding the flaring period (left histogram) is well reproduced by a lognormal law.Flux distribution including the flaring period (right histogram) can be described by the superposition of two lognormal lawsLognormal behaviour in the flaring period confirmed by strong correlations (ρ > 0.80) between excess r.m.s. of flux in small time bins (measurement errors subtracted) and flux.


Strong correlation between flux fluctuations (« excess flux r.m.s. ») and mean flux during the flaring period The flaring state can be described by a stationary random lognormal process → suggests a multiplicative process ...... similar to what is observed in X-rays for X-ray binaries and for Seyfert galaxies. A disk-jet connection ?There is no evidence of cutoff at the highest frequencies (minute.-1) allowed by the sensitivity of H.E.S.S. (doubling timescales on the order of a few minutes)The Fourier spectrum between 10-4 and 10-2 Hz in the flaring regime is well described by a power law of frequency P(ν) = K/να with α = 2.06±0.21 i.e. a red noise spectrum

PKS 2155-304 variability is well described as a red noise (arXiv:1005.3702, A&A accepted)

20 min. light curve segments

80 min. segments


4. Prospects: towards H.E.S.S. II

H.E.S.S. II: a very large telescope (28 m diameter) in the centre of the present array (2012) + camera comprising 2048 pixels30 GeV threshold with the very large telescope80 GeV threshold with the very large telescope + one of the other 4Sensitivity × 2 for E>200 GeV Take advantage of the complementarity of satellites and ground-based instruments for observing variable sources (AGN) or transient events (gamma-ray bursts)

Satellites (large field of view) give the alert in case of a flare Ground-based detectors (large effective area) can measure the light curve with a high temporal resolution e.g. minute-sale variations.


Backup slides


Galactic sources: the identification problem

Supernova Remnants (SNR)Shell-type: γ-ray morphology coincide with radio/X-ray shell:

Shell-type + Molecular cloud: γ-ray emission from contact region between shell and molecular cloud + 1720 MHz OH masers (shock wave in cloud)

Pulsar Wind Nebula (PWN): presence of a pulsar surrounded by an X-ray nebula, γ-ray nebula may be significantly more extended than X-ray nebula

Binary systems: presently 2 High-Mass X-ray Binaries: PSR 1259-63, LS 5039 and one likely candidate HESS J0632+057

Point-like

Modulated emission

Unidentified sourcesSeveral plausible associations (i.e. Galactic Centre)

Stellar winds in star formation regions (OB associations)

No counterparts at other wavelengths: ancient PWN ?


Resolved shell-type supernova remnants

Supernova remnants (SNR’s) are suspected to be the main accelerators of galactic cosmic-rays up to the knee

Diffusive shock acceleration

10% of the mechanical energy of supernovae can account for the injection of CR’s in the Galaxy.

Thin filaments seen in non-thermal (synchrotron) X-rays prove that electrons are accelerated up to 100 TeV and indicate the shock position

H.E.S.S. provided resolved images of several young shell-type SNR’s

RX J1713-3946 RX J0852.0-4622 RCW 86

A&A 464 (2007) 235 ApJ 661 (2007) 236 ApJ 692 (2009) 1500


Supernova remnants interacting with molecular clouds :

the case of W28 A&A 481 (2008) 401

An old (> 35 kyr) supernova remnantRadio shell (white circle)Only the part of the shell coincident with a molecular cloud is detected in gamma-rays :

Only low-energy electrons (emitting in the radio range) remain in most parts of the shellHigh-energy particles are concentrated in the molecular cloud

The interaction between the remnant and the cloud is traced by 1720 MHz OH masers Hadronic scenario is likely : H.E.S.S. and Fermi (arXiv:1005.447) spectra in continuityOverdensity factors of cosmic rays within the clouds (with respect to average density in the Galaxy) of the order on 10 to 20


Supernova remnants interacting with molecular clouds ? New candidates in complex regions ... but ...

CTB 37A / G348.5+0.1 / HESS 1714-3857

• Hadronic scenario : interaction with molecular clouds (OH masers)

• But X-ray non-thermal emission (Chandra) → plausible Pulsar Wind Nebula

W51C / G49.2-0.7 / HESS J1923+141

• Molecular cloud W51B associated with a star-forming region : interaction supported by the presence of OH masers

• But ASCA and Chandra have detected a Pulsar Wind Nebula CXO J192318.5+143035 Koo et al. 2005


Pulsar-Wind NebulaeSome supernova remnants include a pulsar whose strong wind termination shock accelerates particles and thus powers a nebula at the back of the ejecta → synchrotron (radio to X-rays) + inverse Compton = steady signal in very-high-energy γ-rays ≠. periodic signal from the pulsar magnetosphere (as detected by Fermi-LAT at E<10-20 GeV)Some pulsar-wind nebulae are extended (≠ Crab Nebula, unresolved by Cherenkov telescopes), sometimes more than the corresponding X-ray nebula (in older objects). Sometimes, the nebula is displaced with respect to the pulsar (HESS J1825-137, Vela X)In old and extended pulsar-wind nebulae, the γ-ray spectrum steepens away from the pulsar (case of HESS J1825-137) where electrons are « older »

Courtesy of P. Slane

Vela X

HESS J1825-137 Spectral index vs; distance from pulsar

A&A 460 (2006) 365HESS J1825-137


Unidentified sources : ancient pulsar wind nebulae ?

Most of galactic sources detected by H.E.S.S. are pulsar-wind nebulae

Some unidentified sources are likely to be ancient pulsar-wind nebulae (de Jager et al. arXiv:0906.2644)

Magnetic field decreases with time. A weak field → almost no more synchrotron losses → source undetectable in radio or X-rays

Inverse Compton losses dominate → very-high-energy gamma-rays

An unidentified source (HESS J1708-443) in the vicinity of PSR B1706-44 and of the supernova remnant candidate G343.1-2.3 (radio arc)

Region A (pulsar) : no signal found

Region B (circle centered on the radio arc, encompassing the whole radio structure) → 7σ signal (17% of the Crab nebula flux) with a hard spectrum Γ = 2.0 ± 0.1 stat ± 0.2sys


Unidentified sources :Star-forming regions and OB associations ?

Young clusters of massive stars (OB associations) : O and B stars (M>8Msol) are those which end as supernovae → in this environment, particle acceleration may be due to strong stellar winds (presence of Wolf-Rayet stars) Westerlund 2 revisited with 45.5 hours of good quality data

HESS J 1023-575 now detected at 16σ, but Fermi-LAT discovery of the coincident young pulsar PSR J 1022-5746 gives more weight to the interpretation of the source as a pulsar-wind nebula. Intrinsic extension 0.18°±0.02°.Another H.E.S.S. source (SW of the former one) appears clearly wen the energy threshold is increased: HESS J 1026-582 (7σ). Intrinsic extension 0.14°±0.03°. Spectrum harder than that of the first source : Γ = 1.94 ± 0.20stat ± 0.20sys

Westerlund 1, the largest stellar cluster of our Galaxy (6 × 104 solar masses) : with 34 hours of good quality data, H.E.S.S. detects a very extended γ-ray source (~2°) at 17 σ

Westerlund 2

Westerlund 1


NGC 253: Fermi-LAT vs. H.E.S.S.


Centaurus A: Fermi-LAT vs H.E.S.S.


Blazars detected at very high energies are mostly high-frequency-peaked BL Lac’s (HBL)

Blazars are characterized by a two-bump broad-band spectrum E2 dΦ/dE

Synchrotron bumpGamma-ray bump, generally attributed to inverse Compton scattering

Classification according to broad-band spectrumLow-frequency-peaked BL Lac’s (LBL) : synchrotron peak in IR/optical, gamma-ray bump peaks in the GeV range (Fermi-LAT)High-frequency-peaked BL Lac’s (HBL) : synchrotron peak in UV/X, γ-ray bump peaks in the 100 GeV range (Cherenkov telescopes)

3 new blazars detected in the GeV range by AGILE or Fermi-LAT recently detected by H.E.S.S., including one LBLOnly 2 blazars from the H.E.S.S.catalogue have not been detected in the GeV range (1ES 0229+200 and PKS 0548-322)

Fossati et al. 1998


PKS 2155-304 flaring regime Light curves in 4 energy bands (4 nights, July 2006)

arXiv:1005.3702

Photon index vs. time

E>0.2 TeV

0.2 TeV<E<0.35 TeV

0.35 TeV<E<0.6 TeV

0.6 TeV<E


PKS 2155-304 spectral variability (arXiv:1005.3702)

Photon index vs flux:

Dark points correspond to the flaring period


No anomaly in the H.E.S.S. cosmic-ray electron spectrum

First ground-based measurement of cosmic-ray electron spectrum (0.34 TeV - 5 TeV)

Use data far from Galactic Plane and exclude the vicinity of known sources (small contribution of diffuse γ-rays)

Electron/hadron discrimination by machine-learning algorithm → variable ζ : electron peak for ζ =1, hadrons accumulate at low values of ζ

Requiring ζ >0.6 removes 98-99.5% of hadrons.

Hadron background for ζ >0.6 is determined on the basis of proton shower simulations A&A 508 (2009) 561


Indirect search for Dark Matter (DM) γ-rays from DM annihilation

Galactic Centre : HESS J1745-290 but the spectrum is mainly due to an unidentified astrophysical source. (A&A 503 (2009) 817), either the central black hole Sgr A* (but see also A&A 492 (2008) 492) or the pulsar wind nebula G359.95-0.04

DM component cannot be greater than 10% of the source flux → upper limit on <σv> are of the order of 10-24-10-23 cm3s-1 (Phys.Rev.Lett 97 (2006)221102)

M87 centre of the Virgo supercluster: but the variability of the signal (HEGRA,H.E.S.S.,MAGIC) excludes DM as the main source

Nearby dwarf spheroidal galaxies are interesting targets:

High Mass/Luminosity ratio inferred from the radial velocity dispersion of stars

No warm or hot gas, little dust and no cosmic rays

dssrJwithJdE

dN

m

v

dE

dlos )]([)(

8

1 22

00 exp

E

EEI

dE

d

Γ=2.10±0.04sta±0.1sys

E0=15.7±3.4sta±2.5sys

Galactic Centre


H.E.S.S. observation of Dwarf Spheroidal GalaxiesH.E.S.S. observed 3 Dwarf Spheroidal Galaxies (Sagittarius dSph, Carina dSph and Sculptor dSph) and the overdensity Canis Major, the possible remnant of a disrupted dwarf galaxy

No signals → Upper limits (95%) on flux for photon indices between 1.8 and 2.4DM profile models (NFW, core profile) + Particle physics models → Upper limits on <σv> (pMSSM→LSP, Kaluza-Klein →LKP)

Sagittarius dSph (Astropart. Phys. 29 (2008) 55 & 33 (2010) 274)Distance 24 kpc (behind the Galactic Bulge)Φ(E>250 GeV) < 3.6 ×10-12 cm-2 s-1

<σv> < 10-23 – 10-24 cm3 s-1

Sculptor dSph Distance 79 kpcΦ(E>320 GeV) < (3.7 to 4.1) ×10-12 cm-2 s-1

<σv> < 10-22 – 10-21 cm3 s-1

Carina dSphDistance 101 kpcΦ(E>320 GeV) < (1.6 to 2.0) ×10-12 cm-2 s-1

<σv> < 10-22 – 10-21 cm3 s-1

Canis Major overdensity (ApJ 691 (2009) 175)Distance 8 kpc<σv> < 10-22 – 10-21 cm3 s-1

Sculptor dSph, A.Charbonnier for

the H.E.S.S. collaboration., COSPAR 2010

Fermi limits for NFW

B. Degrange ICATPP, October 2010 1 Recent results from H.E.S.S. Bernard Degrange, for the H.E.S.S. collaboration LLR, Ecole polytechnique & CNRS/IN2P3.

Documents