The Very-High-Energy Gamma-Ray Sky and the CTA Observatory

The very-high-energy gamma-ray sky and the CTA Observatory

Jürgen Knödlseder (IRAP, Toulouse) Directeur de Recherche (CNRS) The menu

Light starter I. Why we do gamma-ray astronomy

Assortment of Appezers II. What have we learned so far

Main Dish III. What comes next

Desert IV. Concluding remarks First course

I. Why we do gamma-ray astronomy

A historical introducon The discovery of cosmic rays 1910

1920

1930

1940

1950

1960 Viktor Franz Hess (1912)

1970

1980

1990

2000

2010 The nature of cosmic rays 1910 Charged 1920 parcles! Gamma 1930 rays!

1940

1950

1960

1970

1980

1990 Robert Millikan and Arthur Holly Compton (1931) A hot debate (1932) 2000

2010 Cosmic charged parcles ! 1910

1920

1930

1940 MS Chrisan Huygens 1950 Clay and Berlage (1932) 1960

1970 Geiger counter 1980 4 cm gold bar

1990 Geiger counter

2000 Bothe and Kohlhörster (1929) 2010 Cosmic stac 1910

1920

1930

1940

1950

1960

1970

1980

1990

2000 Karl Jansky (1933) 2010 An ambious amateur 1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

Grote Reber (1944) 2010 … and the ﬁrst radio sky map 1910

1920 Cas A

1930 Cygnus X 1940

1950

1960

1970

1980 Galactic centre 1990

2000

2010 Synchrotron radiaon 1910

1920

1930

1940

1950

1960 Langmuir,Elder,Gurevitsch, 1970 Charleton et Pollock (1948)

1980

1990

2000 Discovery of synchrotron radiaon (1947) 2010 Consequences 1910 Cosmic-ray parcles emit gamma rays! 1920

1930 Hayakawa (1952)

1940 Hutchinson (1952)

1950

1960 Feenberg and Primakoﬀ (1948)

1970

1980 Curvature Radiation 1990

2000

2010 Radhakrishnan and Cooke (1969) Philip Morrison (1958) The dawn of gamma-ray astronomy 1910 or how to observe photons of the highest energies?

1920

1930

1940 Atmosphere barrier Flux barrier 1950

1960

1970

1980 1 photon per cm2/keV per year !

1990 1 photon per cm2/keV per century !

2000

2010 Going into space 1910 or how to overcome the atmosphere barrier

1920

1930

1940

1950 Explorer XI (1962)

OSO 3 (1967) 1960 Comparison of eﬀecve 1970 detecon area (to scale)

1980 COS-B (1975) 1990 EGRET (1991)

2000

2010 Fermi (2008) Observing from ground 1910 or how to overcome the ﬂux barrier (and turn the atmosphere into a detector)

1920

1930

1940

1950

1960

1970

1980

1990

2000

2010 Observing from ground 1910 The Cherenkov technique in a nutshell

1920

1930

1940

1950

1960

1970

1980

1990

2000

2010 Observing from ground 1910 The stereo Cherenkov technique in a nutshell

1920

1930

1940

1950

1960

1970

1980

1990

2000

2010 Observing from ground 1910

1920

1930

1940

1950

1960

1970

1980

1990

2000 Galbraith & Jelley (1953) 2010 Observing from ground 1910

1920

1930

1940

1950 Crimée (1960-1963) Photographie d’une gerbe (1961) 1960

1970

1980

1990

2000

2010 Irlande (1963-1967) Whipple (1968-1990) Observing from ground 1910

1920

1930

1940

1950

1960 HESS (2004-)

1970

1980

1990

2000

2010 MAGIC (2004-) VERITAS (2007-) Observing from ground 1910

1920

1930

1940

1950 MILAGRO (2004-2008) 1960

1970

1980

1990

2000 HAWC (2014-) 2010 Second course

II. What have we learned so far

In the VHE domain (ground-based gamma-ray astronomy) with a strong focus on Galacc science A hundred of gamma-ray sources The source zoo SNRs Binaries Starbursts AGN Pulsars/PWN Jets Shocks Shocks Jets Accreon Diﬀusion Diﬀusion Cosmic Winds rays GRBs

Dark Jets matter Shocks

Intergalactic VHE gamma rays probe medium Parcle acceleraon and propagaon Cosmic rays Highest energy processes in a large variety of objects Intergalacc medium Dark maer and fundamental physics Significance, in full detail Galacc populaons Significance 5 1

) . S . S . E . H (

+ s e v a h C

H.E.S.S. Galacc Plane Survey

, l i e D VHE source distribution Galacc populaons

Galacc latude proﬁle of H.E.S.S. Galacc Plane Survey sources GalacticGalacc populaons TeV source populations

The cosmic-ray spectrum or the quest for the Galacc PeVatron Parcle acceleraon at shocks or the Fermi mechanism

Parcles gain energy by crossing and recrossing the shock The Hillas criterion or what systems can accelerate parcles

Systems below the lines are excluded Supernova remnants as parcle

R. D. Blandford et al. / Nuclearaccelerators Physics B Proceedings Supplement 00 (2014) 1–14 3

Figure 3: Chandra X-ray 4.0 6.0 keV image of the Tycho supernova remnant, showing detailed non-thermal emitting features manifesting the Non-thermal X-ray emission is a clear sign for electron acceleraon to shock fronts (from [16]). ~ 100 TeV (synchrotron emission of electrons in ampliﬁed B ﬁelds)

are not obviously seen in any other spectral band (eg. [32, 33] and [34] for a review). Figure 5 shows the light curve of the biggest flare in April 2011. The flux doubled within td . 8 hours at the rising edges. The spectral energy distribution for a few flares taken near the peak flux level is shown in Figure 6. The emission process is most likely to be synchrotron, requiring electrons/positrons of energy 3 PeV in a 1 mG mag- ⇠ ⇠ netic field. A peak of 400 MeV indicates very e- ⇠ cient acceleration that goes beyond the classical radiation reaction limit in an MHD setting. These flares are not accompanied by changes in the pulsar timing and so presumably originate in the nebula which is many Figure 4: Gamma-ray spectrum of W44 as measured with the Fermi light years in size. The isotropic energy radiated in the LAT, which shows good agreement with pion-decay gamma-ray pro- strongest flare was 1034 J which is equivalent to the ⇠ duction model (from [17]). energy contained within a region about a hundred, not ten, light hours across. Either there is strong relativis- 1.3. Crab Nebula tic beaming or some way must be found to concentrate energy within a small volume (or both). The Crab Nebula has long been our best high energy astrophysics laboratory and many common e↵ects have 1.4. Relativistic Jets first been identified there. It is not disappointing us. Re- cent discoveries include 400 GeV pulsation [28, 29], Fermi and the ACTs have also made dramatic ob- ⇠ rapid secular variation in the total nebular flux [30] and servations of relativistic jets from AGN, GRB and bi- rapid variation of the “inner knot” which may mark the nary sources. Blazars (AGN directed towards us) ex- termination of the wind in the inner nebula [31]. How- hibit variability on timescales that can be as short as few ever the most striking discovery, which may presage a minutes (e.g., [35, 36]). However, we can place a lower new kind of particle acceleration with clear relevance to bound on the radius of emission because the gamma cosmic ray origin, is the discovery of dramatic, 10 hr rays have to avoid pair production as they escape the ⇠ -ray flares which are localized around 400 MeV and near infrared photons. The far more luminous quasars, Hadronic and leptonic emission the pion bump as a disncve feature

Eγ ~ 0.1 Ep R. D. Blandford et al. / Nuclear Physics B Proceedings Supplement 00 (2014) 1–14 3

Figure 3: Chandra X-ray 4.0 6.0 keV image of the Tycho supernova remnant, showing detailed non-thermal emitting features manifesting the shockIndicaons for a pion bump fronts (from [16]). but not a Galacc PeVatron are not obviously seen in any other spectral band (eg. [32, 33] and [34] for a review). Figure 5 shows the light curve of the biggest flare in April 2011. The flux doubled within td . 8 hours at the rising edges. The spectral energy distribution for a few flares taken near the peak flux level is shown in Figure 6. The emission process is most likely to be synchrotron, requiring electrons/positrons of energy 3 PeV in a 1 mG mag- ⇠ ⇠ netic field. A peak of 400 MeV indicates very e- ⇠ cient acceleration that goes beyond the classical radiation reaction limit in an MHD setting. These flares are not accompanied by changes in the pulsar timing and so presumably originate in the nebula which is many Figure 4: Gamma-ray spectrum of W44 as measured with the Fermi light years in size. The isotropic energy radiated in the LAT, which shows good agreement with pion-decay gamma-ray pro- strongest flare was 1034 J which is equivalent to the ⇠ duction model (from [17]). energy contained within a region about a hundred, not ten, light hours across. Either there is strong relativis- 1.3. Crab Nebula tic beaming or some way must be found to concentrate energy within a small volume (or both). The Crab Nebula has long been our best high energy astrophysics laboratory and many common e↵ects have 1.4. Relativistic Jets first been identified there. It is not disappointing us. Re- cent discoveries include 400 GeV pulsation [28, 29], Fermi and the ACTs have also made dramatic ob- ⇠ rapid secular variation in the total nebular flux [30] and servations of relativistic jets from AGN, GRB and bi- rapid variation of the “inner knot” which may mark the nary sources. Blazars (AGN directed towards us) ex- termination of the wind in the inner nebula [31]. How- hibit variability on timescales that can be as short as few ever the most striking discovery, which may presage a minutes (e.g., [35, 36]). However, we can place a lower new kind of particle acceleration with clear relevance to bound on the radius of emission because the gamma cosmic ray origin, is the discovery of dramatic, 10 hr rays have to avoid pair production as they escape the ⇠ -ray flares which are localized around 400 MeV and near infrared photons. The far more luminous quasars, Evoluon of the SNR accelerator HESSHESS Indicaons for a J1641-463: J1641-463: PeVatron HESS J1641-463 p-pleptonic model modelfits data doesn't better fit → well Emax (Klein-Nishina > 100 TeV (99% effect) CL)

Assuming a PL for Kelner+06 parameterization primary electron spectrum RX J1713 for Assuming a PL withcomparison Γ = 2.1 Requires electron primary proton spectrum cutoff > 700 TeV (99% CL)

to fit measured VHE One of the highest Emax every spectrum inferred from VHE data 4

Hundreds of TeV electrons measured luminosity (d = 11 kpc) 1

34 -1 ) suffer severe losses in L = 4 ¥ 10 erg s . γ S .

amplified B-field S cooling time . E

15 -3 -1 . t ≈ 5 ¥ 10 (n / cm ) s H π0 (

No break in photon spectrum + i k s

total energy SN → hadronic acc. w

o Strongly disfavored 50 -3 -1 m Wp = Lγ tπ0 ≈ 10 (n / cm ) erg a r

scenario b A LETTER doi:10.1038/nature17147

Acceleration of petaelectronvolt protons in the Galactic Centre HESS Collaboration*

Galactic cosmic rays reach energies of at least a few petaelectronvolts1 more than 100 GeV) emission of the central molecular zone. This region (of the order of 1015 electronvolts). This implies that our Galaxy surrounding the Galactic Centre contains predominantly molecular gas contains petaelectronvolt accelerators (‘PeVatrons’), but all proposed and extends (in projection) out to radius r ≈ 250 pc at positive Galactic models of Galactic cosmic-ray accelerators encounter difficulties longitudes and r ≈ 150 pc at negative longitudes. The map of the central at exactly these energies2. Dozens of Galactic accelerators capable molecular zone as seen in VHE γ-rays (Fig. 1) shows a strong (although of accelerating particles to energies of tens of teraelectronvolts not linear; see below) correlation between the brightness distribution (of the order of 1013 electronvolts) were inferred from recent γ-ray of VHE γ -rays and the locations of massive gas-rich complexes. This observations3. However, none of the currently known accelerators— points towards a hadronic origin of the diffuse emission7, where the not even the handful of shell-type supernova remnants commonly γ-rays result from the interactions of relativistic protons with the ambi- believed to supply most Galactic cosmic rays—has shown the ent gas. The other important channel of production of VHE γ -rays is characteristic tracers of petaelectronvolt particles, namely, power- the inverse Compton (IC) scattering of electrons. However, the severe law spectra of γ-rays extending without a cut-off or a spectral break radiative losses suffered by multi-TeV electrons in the Galactic Centre to tens of teraelectronvolts4. Here we report deep γ-ray observations region prevent them from propagating over scales comparable to the with arcminute angular resolution of the region surrounding the size of the central molecular zone, thus disfavouring a leptonic origin of Galactic Centre, which show the expected tracer of the presence the γ-rays (see discussion in Methods and Extended Data Figs 1 and 2). of petaelectronvolt protons within the central 10 parsecs of the The location and the particle injection rate history of the cosmic-ray Galaxy. We propose that the supermassive black hole Sagittarius accelerator(s) responsible for the relativistic protons determine the A* is linked to this PeVatron. Sagittarius A* went through active spatial distribution of these cosmic rays which, together with the gas phases in the past, as demonstrated by X-ray outbursts5 and an distribution, shape the morphology of the central molecular zone outflow from the Galactic Centre6. Although its current rate of seen in VHE γ-rays. Figure 2 shows the radial profile of the E ≥ 10 TeV particle acceleration is not sufficient to provide a substantial cosmic-ray energy density wCR up to r ≈ 200 pc (for a Galactic Centre contribution to Galactic cosmic rays, Sagittarius A* could have distance of 8.5 kpc), determined from the γ -ray luminosity and the plausibly been more active over the last 106–107 years, and therefore amount of target gas (see Extended Data Tables 1 and 2). This high should be considered as a viable alternative to supernova remnants energy density in the central molecular zone is found to be an order of as a source of petaelectronvolt Galactic cosmic rays. magnitude larger than that of the ‘sea’ of cosmic rays that universally The large photon statistics accumulated over the last 10 years of fills the Galaxy, while the energy density of low energy (GeV) cosmic observations with the High Energy Stereoscopic System (HESS), rays in this region has a level comparableIndicaons for a to it8. This requires the pres- PeVatron together with improvements in the methods of data analysis, allow for ence of one or more accelerators of multi-TeV particles operating in LETTER RESEARCH a deep study of the properties of the diffuse very-high-energy (VHE; the central molecular zone. Galacc Centre

10–10 30 160.0 +00.6 a +00.4 b +00.4 20 61.7 +00.2 ) +00.2 23.0 Counts per pixel –3

+00.0 +00.0 7.8

eV cm 10 Sgr A* Sgr A* 10–11 –3 1/r –00.2 0 –00.2 (1 1.9 ) –1

V) 6.0 × local cosmic-ray density s

–00.4 Te Galactic latitude (degrees) –2 –00.4 10 –0.5 cm ≥ –00.6 ( 1/r 2 CR eV w –1.4 (T 01.0 00.5 00.0 359.5 359.0 00.0 359.5 x u

f –12 Galactic longitude (degrees) Galactic longitude (degrees) 10 ×

30 2 Figure 1 | VHE γ-ray image of the Galactic Centre region. The colour2 molecular gas, as traced by its CS line emission . Black star, location of E scale indicates counts per 0.02° × 0.02° pixel. a, The black lines outline Sgr A* . Inset Diﬀuse VHE emission without any (bottom left), simulation of a point-like source. The part of the regions used to calculate the cosmic-ray energy density throughout the image shown boxed is magnified in b. b, Zoomed view of the inner Diffuse emission (×10) 0 20∼ 70 pc40 and thecut-oﬀ 60 contour80 of the100 region120 used to140 extract160 the spectrum180 of200 the the central molecular zone. A section of 66° is excluded from the annuli Model (best ft): diffuse emission (see Methods). White contour lines indicate the density distribution of diffuse emission. Projected distance (pc) 68% Radial distribuon compable Model: diffuse emission E cut,p = 2.9 PeV Lists of participants and their affiliations appear at the end of the paper. 90% * Figure 2 | Spatial distribution of the cosmic-ray density versus Model: diffuse emission E cut,p = 0.6 PeV with connuous proton injecon –13 95% projected distance from Sgr A*. The vertical and horizontal error bars 10 Model: diffuse emission E cut,p = 0.4 PeV 476 | NATURE | VOL 531 | 24 MARCH 2016 show the 1σ statistical plusover last 1000 years systematic errors and the bin size, respectively. HESS J1745–290 © 2016 MacmillanFits to Publishers the data Limited. of a 1/Allr rights (red reserved line, χ2/d.o.f. = 11.8/9), a 1/r2 (blue line, χ2/ d.o.f. = 73.2/9) and a homogeneousSupermassive black hole? (black line, χ2/d.o.f. = 61.2/9) cosmic- 1 10 ray density radial profile integrated along the line of sight are shown. The Energy, E (TeV) best fit of a 1/rα profile to the data is found for α = 1.10 ± 0.12 (1σ). The 1/r radial profile is clearly preferred for the HESS data. Figure 3 | VHE γ-ray spectra of the diffuse emission and HESS J1745−290. The y axis shows fluxes multiplied by a factor E2, where E is the −2 −1 If the accelerator injects particles (here we consider protons through- energy on the x axis, in units of TeV cm s . The vertical and horizontal σ out) at a continuous rate, Q! ()E , the radial distribution of cosmic rays error bars show the 1 statistical error and the bin size, respectively. Arrows p represent 2σ flux upper limits. The 1σ confidence bands of the best-fit in the central molecular zone, in the case of diffusive propagation, is QE! () spectra of the diffuse and HESS J1745−290 are shown in red and blue described9 as w ()Er,,t = p erfc(r/r ), where D(E) and r are CR 4πDE( )r diff diff shaded areas, respectively. Spectral parameters are given in Methods. The γ the diffusion coefficient and radius, respectively. For timescales t red lines show the numerical computations assuming that -rays result from the decay of neutral pions produced by proton–proton interactions. The smaller than the proton–proton interaction time (tpp ≈ 4 3 −1 −3 fluxes of the diffuse emission spectrum and models are multiplied by 10 to 5 × 10 (n/10 ) yr, where n is the density of the hydrogen gas in cm ), visually separate them from the HESS J1745−290 spectrum. the diffusion radius is rdiff ≈ 4DE( )t . Thus, at distances r < rdiff, the proton flux should decrease as ∼ 1/r provided that the diffusion coef- spectrum implies that the spectrum of the parent protons should extend ficient does not vary much throughout the central molecular zone. The to energies close to 1 PeV. The best fit of a γ-ray spectrum from neutral measurements clearly support the wCR(r) ∝ 1/r dependence over the pion decay to the HESS data is found for a proton spectrum following entire central molecular zone region (Fig. 2) and disfavour both a pure power law with an index of ∼ 2.4. We note that pp interactions 2 wCR(r) ∝ 1/r and wCR(r) ∝ constant profiles (the former is expected if of 1 PeV protons could also be studied by the observation of emitted cosmic rays are advected in a wind, and the latter in the case of a single neutrinos or X-rays from the synchrotron emission of secondary elec- burst-like event of cosmic-ray injection). The 1/r profile of the cos- trons and positrons (see Methods and Extended Data Figs 3 and 4). mic-ray density up to 200 pc indicates a quasi-continuous injection of However, the measured γ -ray flux puts the expected fluxes of neutri- protons into the central molecular zone from a centrally located accel- nos and X-rays below or at best close to the sensitivities of the current erator on a timescale ∆ t exceeding the characteristic time of diffusive instruments. Assuming a cut-off in the parent proton spectrum, the escape of particles from the central molecular zone, that is, ∆t ≥ tdiff ≈ corresponding secondary γ-ray spectrum deviates from the HESS data R2/6D ≈ 2 × 103(D/1030)−1 yr, where D (in cm2 s−1) is normalized to at 68%, 90% and 95% confidence levels for cut-offs at 2.9 PeV, 0.6 PeV the characteristic value of multi-TeV cosmic rays in the Galactic disk10. and 0.4 PeV, respectively. This is the first robust detection of a VHE In this regime the average injection rate of particles is found to cosmic hadronic accelerator which operates as a source of PeV particles ! 37 30 −1 be Qp(≥10 TeV4) ≈ × 10 (D/)10 erg s . The diffusion coefficient (a ‘PeVatron’). itself depends on the power spectrum of the turbulent magnetic field, Remarkably, the Galactic Centre PeVatron appears to be located which is unknown in the central molecular zone region. This intro- in the same region as the central γ -ray source HESS J1745− 290 duces an uncertainty in the estimates of the injection power of relativ- (refs 11–14). Unfortunately, the current data cannot provide an answer istic protons. Yet, the diffusive nature of the propagation is constrained as to whether there is an intrinsic link between these two objects. The by the condition R2/6D ≫ R/c. For a radius of the central molecular point-like source HESS J1745−290 itself remains unidentified. Besides zone region of 200 pc, this implies D ≪ 3 × 1030 cm2 s−1, and, conse- Sgr A* (ref. 15 ), other potential counterparts are the pulsar wind nebula ! 38 −1 18 quently, Qp ≪ 12. × 10 ergs . G 359.95− 0.04 (refs 16, 17) and a spike of annihilating dark matter . The energy spectrum of the diffuse γ -ray emission (Fig. 3) has been Moreover, it has also been suggested that this source might have a extracted from an annulus centred at Sagittarius (Sgr) A* (see Fig. 1). diffuse origin, peaking towards the direction of the Galactic Centre The best fit to the data is found for a spectrum following a power law because of the higher concentration there of both gas and relativistic extending with a photon index of ∼ 2.3 to energies up to tens of TeV, particles15. In fact, this interpretation would imply an extension of the without a cut-off or a break. This is the first time, to our knowledge, spectrum of the central source to energies beyond 10 TeV, which how- that such a γ -ray spectrum, arising from hadronic interactions, has ever is at odds with the detection of a clear cut-off in the spectrum of been detected. Since these γ-rays result from the decay of neutral pions HESS J1745−290 at about 10 TeV (refs 19, 20; Fig. 3). Yet the attractive produced by pp interactions, the derivation of such a hard power-law idea of explaining the entire γ-ray emission from the Galactic Centre by

Protons leaking out of SNR illuminate surrounding molecular clouds

Indicaon for a smaller diffusion coefficient than Galacc average, maybe as a result of turbulence Cosmic-ray illuminaon another case for proton diffusion

Requires good knowledge about gas distribuon near accelerators Enables measuring of diﬀusion parameters ANRV284-AA44-02 ARI 28 July 2006 13:47

Pulsar Wind Nebulae the prevailing Galacc VHE source class

Figure 2 The end products of a massive (>8 (a) A deep Chandra X-rayMsolimage) star: a compact object and a SNR of the composite SNR G21.5–0.9 (Matheson & Saﬁ-Harb Pulsar produces a wind of electrons forming a nebula (PWN). Electrons further 2005). A circular supernova remnant (SNR) of diameter 5 surrounds a symmetric pulsar ≈ ′ wind nebula (PWN) of diameter 1.5, with the young pulsar J1833-1034 at the center (Gutpa accelerated at wind terminaon shock. ≈ ′ et al., 2005; Camilo et al., 2006). The central location of the pulsar and PWN and the symmetric appearance of the PWN and SNR both argue for a relatively unevolved system in which the PWN expands freely and symmetrically into the unshocked interior of the SNR. (b) A schematic diagram of a composite SNR showing the swept-up interstellar medium shell, hot and cold ejecta separated by the reverse shock, and the central pulsar and its nebula. The expanded PWN view shows the wind termination shock. Note that this diagram does not correspond directly to G21.5–0.9, in that a signiﬁcant reverse shock has probably yet to form in this young SNR.

have been recently discussed by van der Swaluw, Downes & Keegan (2004), Melatos (2004), and Chevalier (2005). The outline of this review is as follows: in Section 2 we explain the basic ob- servational properties of pulsars and their nebulae; in Section 3 we review current understanding of the evolutionary sequence spanned by the observed population of PWNe; in Section 4 we discuss observations of PWNe around young pulsars, which by OBSERVATOIRE MIDI PYRENEES on 03/07/07. For personal use only. represent the most luminous and most intensively studied component of the population; in Section 5 we consider the properties of the bow shocks produced by high Annu. Rev. Astro. Astrophys. 2006.44:17-47. Downloaded from arjournals.annualreviews.org velocity pulsars; and in Section 6 we brieﬂy describe other recent and interesting results in this ﬁeld.

2. OVERALL PROPERTIES 2.1. Pulsar Spin Down

Because a pulsar’s rotational energy, Erot , is the source for most of the emission seen from PWNe, we ﬁrst consider the spin evolution of young neutron stars.

20 Gaensler Slane · 4 1

) .

Studying a possible S . S . E

e.g. .

pulsar wind nebulae for H (

+ i

E-dep. morphology... k s w o m a r b A 4 1

) .

Studying a possible S .

Pulsar Wind Nebulae S . E

e.g. .

pulsar wind nebulae for H (

+ i

E-dep. morphology... k s

HESS J1640-465 w o m a r b A

High-electrons loose energy more quickly than low-energy electrons, leading to a decrease of PWN size with increasing energy. Study of electron acceleraon and cooling in a complex environment.

HESS J1640-465

Gamma-ray binaries and related systems 11

10 Guillaume Dubus

Gamma-ray binaries

LS 5039 LS I +61°303

] 1.6 -1 s -2 1.4

1.2 ph cm -6 1

Flux [ 10 0.8

0.6

0.4 Fermi, after flux change 0.2 Fermi, 8 months

0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Phase Fig. 4 Observations of LS 5039. From top to bottom: spectral energy distribution adapted from Dubus (2006b); X-ray modulation (reproduced byGamma-ray sources exhibing an orbital modulaon, all know examples permissionFig. of5 Observations the AAS from of Kishishita LS I +61 et303. al. 2009); From GeV top to bottom: spectral energy distribution (this work); one modulation (reproduced by permission of the AAS fromexampleFermi of/LAT the collaboration periodic radio et and al. 2009b); X-ray outburst TeV mod- that occur once per orbit (reproduced by permission of the ulation (reproduced with permission from H.E.S.S.associated with high-mass X-ray binaries (O or B primary stars) collaborationAAS from Harrison et al. 2005a et al.c ESO). 2000) ; GeV modulation (reproduced by permission of the AAS from Hadasch Gamma-ray binaries = most radiated power emied in gamma rays et al. 2012); TeV modulation (reproduced by permission of the AAS from VERITAS collaboration et al. 2011b). Periastron passage is at = 0.23. spectrum: a power-law with a photon index 2.7 and a normalisation at 1 TeV VHE ⇡ of 1.3 10 12 TeV 1cm 2 s 1 (H.E.S.S. collaboration et al. 2005b, 2009b, 2013) ⇥ phases 0.5 0.9, peaking at 0.6-0.8 (MAGIC collaboration et al. 2009a, 2006;   LS 5039 was detected in the HESS GalacticVERITAS Plane survey collaboration (H.E.S.S. et collaborational. 2008, 2011b). Later observations then failed to detect et al. 2005a). A Lomb-Scargle periodogramthe source,of the VHE until lightcurve late 2010, gives when a period the source was detected at 0.0 0.1 (VER- of 3.90678 0.0015 days that corresponds to the orbital period determined indepen-   ± ITAS collaboration et al. 2011b). The 26.5 day orbital period is long enough and dently using radial velocity measurements (H.E.S.S. collaboration et al. 2006b). The close enough to the lunar cycle to make repeated, homogeneous sampling of the or- minimum is close to superior conjunction or to periastron (the two phases are separated by only = 0.057). Maximum fluxbital occurs lightcurve around inferior di cult conjunction compared (Fig. to 4). LS 5039. The average spectra reported by both Spectral variability is detected between superiorcollaborations and inferior are compatible conjunction (“SUPC”within systematic errors. The best fit is a power-law 0.45 and 0.9, “INFC” 0.45 with0.9). a At photon INFC, index the bestVHE fit spectrum2.4 to is 2.7, a and a normalisation at 1 TeV from 2.4 to    12 1 2 1 ⇡ power law with 1.8 and an exponential2.9 10 cuto TeV↵ atEcm8.s7 TeV.(MAGIC At SUPC, collaboration et al. 2008, 2012b; VERITAS col- VHE ⇡ ⇥ c ⇡ the source is fainter and best described bylaboration a single et power-law al. 2009a, with 2011b, a softer Fig. index 19). 2.5. The average normalisation at 1 TeV is 1.8 10 12 TeV 1cm 2 s 1 (see VHE ⇡ ⇥ Fig. 19). There is no report of long term changes in the orbit-averaged flux. HESS J0632+057 was detected by the HESS Galactic Plane survey (H.E.S.S. collaboration et al. 2007). The VHE detections cluster around 0.2 0.4 when folded on LS I +61303 was detected by the MAGIC and VERITAS collaborations. Again,   the 315 5 d period (Fig. 6, Bongiorno et al. 2011; Bordas & Maier 2012). In hind- the VHE emission is tied to the orbital motion. However,± unlike LS 5039, the orbital phases of VHE detections have variedsight, considerably the initial with VHE epoch detection (Fig. 5). was Early lucky since follow-up observations by IACTs observations (Oct. 2005-Jan 2008) indicatedfailed that repeatedly VHE emission to re-detect was confined the source to before the long orbital period was identified in the X-ray lightcurve (VERITAS collaboration et al. 2009b; MAGIC collaboration et al. 2012a; Ong 2011). This the only gamma-ray binary that can be observed by Gamma-ray binaries

PWN scenario Micro quasar scenario

PWN interacng with stellar wind of Jets are launched from an accreon O or B star; gamma-ray emission from disk, with gamma-ray emission arising pulsar, PWN or wind-PWN shock from the corona of the accreon disk, within the jet, or at terminaon shock Gamma-ray binaries Gamma-ray binaries and related systems 19 or Pulsar Wind Nebulae in disguise

Radio images of LS I +61°303 show a nebula with a low-energy tail poinng Fig. 9 Maps of the 8 GHz radio emission from LS I +61303 obtained in consecutive VLBA observations covering an orbit of the binary. The above is a montage with the maps (resolution 1.1 1.5 mas) organized away from the primary (B0V star), taken as evidence for the impact of the ⇥ by phase around an exaggerated orbit (real size 0.3 mas). Figure reprinted with permission from Dhawan stellar wind from the primary. et al. (2006).

LS 5039 has persistent radio emission with < 30% variability around a mean level 30 mJy at 2 GHz (Ribo´ et al. 1999). The⇠ radio ﬂux is not known to vary on the ⇡ orbital period. The spectrum is a power-law of spectral index ↵ 0.5 breaking to ⇡ an absorbed, optically thick spectrum below 1 GHz with ↵ 0.75 (Godambe et al. ⇡ 2008; Bhattacharyya et al. 2012). The radio emission was ﬁrst resolved by Paredes et al. (2000) on scales of 1–6 mas ( 2.5 AU). Further observations found that the radio morphology changes with orbital phase and is stable from orbit to orbit (Ribo´ et al. 2008; Moldon´ et al. 2012a). Emission on larger scales (60–300 mas) has also been reported (Paredes et al. 2002).

LS I +61303 ’s 26.5 day orbital period is best determined from its regular radio outbursts (Fig. 5). An additional period at 1667 8 days appears in the radio lightcurve ± accumulated since the 1980s (Gregory 2002). This superorbital period is also present in HE gamma-rays, X-rays, and optical. The peak phase ( 0.4 to 1.0), maximum ⇡ intensity (50–300 mJy), and amplitude (factor 2-10) of the radio outburst depend on the superorbital cycle. Taking this into account, Chernyakova et al. (2012) noticed that the radio peak follows the X-ray peak after a delay of 0.2 (see also Harri- ⇡ son et al. 2000). The spectrum varies from optically thick to thin indices during the outburst (↵ 0.4 to ↵ 0.5, Ray et al. 1997; Massi & Kaufman Bernado´ 2009). ⇡ ⇡ Dhawan et al. (2006) presented VLBI observations throughout an orbit (Fig. 9). The radio source is resolved on mas scales (2 AU), elongated by 7 mas, the position ⇡ angle changing with , with lower frequency emission occurring further away along The Large The LMC inMagellanic VHE γ-rays: Cloud Recap 4 1

) . S . S . E . H (

+ i k s w o

m a r b A Black hole parcle acceleraon convenonal picture is shock acceleraon in jets

H.E.S.S. Aharonian et al. (2007)

PKS 2155-304

Convenonal picture challenged by observed rapid me variability (minutes) Size of the emission region smaller than black hole Schwarzschild radius (diﬃcult to explain in jet shock models) Parcle acceleraon by magnec reconnecon? (like in our Sun) Probes of the Universe pair creation AGN (with halo) e+/e- inverse Compton CMB EBL

-14 10 G e+/e-

-15 10 G e+/e-

Gamma rays allow to probe • intergalacc magnec ﬁelds • intergalacc (infrared) radiaon ﬁeld Third course

III. What comes next

The Cherenkov Telescope Array

SST (!4m)

MST (!12m) LST (!23m)

2 sites (north & south) 3 telescope size classes About 120 telescopes in total CTA Sites

SST (!4m)

MST (!12m) LST (!23m) All sky coverage

South + North >60° zenith 45°-60° 30°-45°

SST (!4m)

MST (!12m) LST (!23m)

South North An open observatory

SST (!4m)

MST (!12m)

The CTA Observatory will provide support to non-expert users Proposal preparaon & submission tools (TAC evaluaon) Calibrated, reconstructed & reduced event data (FITS) Soware to analyse data (Fermi-LAT like) User documentaon Help Desk, Knowledge, Training CTA Consorum

SST (!4m)

MST (!12m)

CTA Consorum members come from CTA Consorum members status 32 countries 1308 members 88 pares 438 FTE 202 instutes Large size telescopes

Science drivers Lowest energies (< 200 GeV) Transient phenomena DM, AGN, GRB, pulsars SST (!4m) Characteriscs Parabolic design 23 m diameter 370 m2 effecve mirror area 28 m focal length 1.5 m mirror facets MST (!12m) 4.5° field of view 0.11° PMT pixels acve mirror control Carbon-fibre arch structure (fast repoinng)

Array layout South site: 4 LST North site: 4 LST

Status Some elements prototyped First full telescope under construcon on La Palma Mid size telescopes

Science drivers Mid energies (100 GeV – 10 TeV) DM, AGN, SNR, PWN, binaries, starbursts, EBL, IGM SST (!4m) Characteriscs Modified Davies-Coon design 12 m diameter 90 m2 effecve mirror area 1.2 m mirror facets MST (!12m) 16 m focal length 8° field of view 0.18° PMT pixels

Array layout South site: 25 MST North site: 15 MST

Status Telescope prototyped (Berlin-Adlershof) Prototype cameras under construcon (2 types: NectarCAM & FlashCam) Mid size telescopes (US opon) Characteriscs Schwarzschild-Couder design 9.7 m primary diameter 5.4 m secondary diameter 40 m2 eﬀecve mirror area SST (!4m) 5.6 m focal length 8° ﬁeld of view 0.07° PMT pixels

Array layout South site: 24 SCT MST (!12m) North site: -

Status Prototype telescope, including camera, under construcon on VERITAS site

Small size telescopes SST 1M ASTRI GCT

SST (!4m)

MST (!12m)

Science drivers Characteriscs Status Highest energies (> 5 TeV) 6.0 - 8.5 m2 effecve mirror area Prototype telescopes built Galacc science, PeVatrons 2.2 - 5.6 m focal length Camera prototypes under 8.6° - 9.6° field of view construcon (first CHEC prototype Array layout 0.16° - 0.24° SiPM pixels build and tested on GCT) South site: 70 SST North site: - Expected performance

Sensivity gain • access VHE populaons across enre Galaxy • sample fast variability (AGN, GRB) SST (!4m)

FoV > 8° • measure diffuse emissions • efficient survey of large fields MST (!12m)

Arcmin angular resoluon • resolve extended sources (SNR, starbursts)

Broad energy coverage • < 100 GeV to reach higher redshis • > 10 TeV to search for PeVatrons Time domain astronomy CTA as a transient factory Huge advantage over Fermi-LAT in energy range of overlap for seconds to week mescale phenomena

SST (!4m)

Example: PKS 2155-304 PKS 2155-304 ﬂare, sampled with 7.5 sec me resoluon MST (!12m) CTA calendar

Dates are only Crical Design Review, Site Decision indicave!

SST (!4m) Headquarters decision, Engagement of Funding Agencies

Construcon MST (!12m)

Early Science Key Science Projects Open Program

2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 … CTA Key Science

Cosmic Parcle Acceleraon How and where are parcles accelerated? How do they propagate? What is their impact on the environment?

Probing Extreme Environments Processes close to neutron stars and black holes? Processes in relavisc jets, winds and explosions? Exploring the intergalacc space

Physics froners – beyond the Standard Model What is the nature of Dark Maer? How is it distributed? Is the speed of light a constant for high-energy photons? Do axion-like parcles exist? Space-based gamma-ray astronomy post Fermi e-ASTROGAM An Advanced Compton telescope 100 keV – 100 MeV

Topics • Low-energy cosmic rays • Compact objects • Gamma-ray bursts • Nucleosynthesis • Anmaer

Prototypes (balloon)

MEGA

Principle COSI TIGRE COSI a ultra-long duraon balloon payload for MeV astronomy

12 Germanium strip detectors

Science objecves • Origin of galacc positrons (511 keV line) • Polarizaon of Gamma Ray Bursts • Nucleosynthesis in the Galaxy COSI a ultra-long duraon balloon payload for MeV astronomy COSI

Horizon camera

… some interesng science ahead Today: 29 days ﬂying Forth course

IV. Concluding remarks

Conclusion The domain of ground-based gamma-ray astrophysics has seen a revoluon during the last decade • starng from a handful of sources we now know more than 100 very-high- energy gamma-ray emiers • from the inial quest for the Galacc source of cosmic rays has emerged a new observaonal window allowing the study of the most extreme and powerful accelerators in the Universe • … with the byproduct of probing the Universe and providing a new tool to unveil the dark maer of our Universe

The next observatory to explore the very-high-energy Universe will be the Cherenkov Telescope Array. CTA will be a major research infrastructure for high-energy astronomy in the next decades.

Also the space-based exploraon of the gamma-ray sky will connue, with as next step the exploraon of the MeV domain, unveiling the low-energy cosmic- ray content of the Universe, and probing compact objects, GRBs, nucleosynthesis and anmaer.