Constraining gamma-ray propagation on cosmic distances
SSF2A 2013
L. Cambr´esy, F. Martins, E. Nuss and A. Palacios (eds)
CONSTRAINING GAMMA-RAY PROPAGATION ON COSMIC DISTANCES
J. Biteau , Abstract.
Studying the propagation of gamma rays on cosmological distances encompasses a variety ofscientific fields, focusing on diffuse radiation fields such as the extragalactic background light, on the probeof the magnetism of the Universe on large scales, and on physics beyond the standard models of cosmologyand particle physics. The measurements, constraints and hints from observations of gamma-ray blazarsby airborne and ground-based instruments are briefly reviewed. These observations point to gamma-raycosmology as one of the major science cases of the Cherenkov Telescope Array, CTA.Keywords: Cosmology: miscellaneous, Cosmic background radiation, Intergalactic medium, Gamma rays:galaxies, BL Lacertae objects: general
Gamma-ray astronomy at very high energies (VHE, 30 GeV < E <
30 TeV) is driven by ground-based imag-ing atmospheric Cherenkov telescopes, whose current major representatives are
H.E.S.S. (Namibia),
MAGIC (Canaries islands) and
VERITAS (Arizona). Together with airborne instruments observing the gamma-ray skyat high energies (HE, 300 MeV < E <
300 GeV), such as
Fermi -LAT , they can now probe and constrain thepropagation on billion-lightyear distances of the highest-energy photons produced and detected in the Universe.The observational field called gamma-ray cosmology encompasses the study of the extragalactic backgroundlight (EBL), an optical-infrared radiative component born with the first stars and that kept on growing sincethen, of the intergalactic magnetic field, which fills the voids between large scale structures and whose originremains unidentified, and of physics beyond the standard models of cosmology and particle physics, possiblyrevealing hints on dark matter candidates and on the laws of nature at the Planck scale.The science case of gamma-ray cosmology is discussed in Sec. 1. A wide-spectrum but non-exhaustive listof current constraints, hints and measurements in the field is exposed in Sec. 2. Possible refinements in theanalyses are discussed in Sec. 3, together with the great expectations from the next generation instrument, theCherenkov Telescope Array (CTA).
Gamma-ray cosmology could be considered as an established domain, with seminal theoretical studies that canbe traced back to the ’60s (Nikishov 1962; Jelley 1966; Gould & Schr´eder 1967), but the first powerful enoughextragalactic accelerators to be used as probes were detected in the beginning of the ’90s. These strong sourcesare discussed in the following subsection together with the science case addressed.
Since the groundbreaking detection of Mrk 421 (Punch et al. 1992) by the first generation ground-based instru-ments Whipple, the number of detected VHE extragalactic sources has exponentially grown, up to almost 60sources in mid 2013. CAT and HEGRA, the major second generation instruments, contributed to the effort,yielding a handful of sources in the beginning of the 2000s. As shown in Fig. 1, left, the third generationinstruments
H.E.S.S. , MAGIC and
VERITAS have revolutionized the field with the discovery of almost 50sources in less than ten years, up to redshifts z > .
6, as shown in Fig. 1, right. Santa Cruz Institute for Particle Physics, Department of Physics, University of California at Santa Cruz, Santa Cruz, CA95064, USA Laboratoire Leprince-Ringuet, Ecole polytechnique, CNRS/IN2P3, F-91128 Palaiseau, Francec (cid:13)
Soci´et´e Francaise d’Astronomie et d’Astrophysique (SF2A) 2013 a r X i v : . [ a s t r o - ph . C O ] O c t
38 SF2A 2013
Year1990 1995 2000 2005 2010 o f T e V ex t r a g a l ac t i c s ou r ces Plot updated in October 2013
Data from TeVCat z redshift 0 0.1 0.2 0.3 0.4 0.5 0.6 o f T e V ex t r a g a l ac t i c s ou r ces Plot updated in October 2013
Data from TeVCat
Fig. 1. Left:
Number of detected extragalactic emitter of VHE gamma rays as a function of the year of discovery.
Right:
Distribution of the redshifts of the current population of extragalactic VHE emitters with a constrained distance.
Datafrom TeVCat http://tevcat.in2p3.fr . Besides the nearby starburst galaxies NGC 253 and M 82, the VHE extragalactic population is exclusivelycomposed of active galactic nuclei (AGN). These objects are thought to host a supermassive black hole, possiblyas heavy as a billion solar masses, fed by an accretion disk, with the emission of relativistic jets on each sideof the disk for the brightest radio-emitters (see Antonucci 1993; Urry & Padovani 1995, for a unified pictureof AGN). With a typical opening angle of few degrees, these jets, which are composed of magnetic fields,accelerated particles and beamed radiation, can either be seen edge on, as for the nearby radio galaxies M 87,Centaurus A and NGC 1275, or they can be closely aligned with the line of sight. In the latter case, the AGN iscalled a blazar and it appears particularly bright and energetic due to the relativistic beaming and the Dopplershift of the energy of the photons (see e.g. Ghisellini 2013, for a recent discussion on gamma-ray blazars).Blazars constitute the vast majority of detected AGN at VHE and can be divided in flat-spectrum radioquasars (3 detected at VHE), which exhibit strong emission lines, and BL Lac objects (49 detected at VHE),with weak or absent lines. The broad band emission of blazars is composed of a radio to UV/X-ray component,generally attributed to the synchrotron emission of relativistic electrons in the jet, and of a gamma-ray compo-nent, whose origin remains debated. The simplest radiative models attribute it to the Comptonization of thesynchrotron field by the very electrons that generated it. This synchrotron self-Compton model (Band & Grind-lay 1985) works fairly well for the largest subclass of BL Lac objects, the high-synchrotron-frequency-peakedobjects (HSP, sometimes HBL in the literature), whose low-energy components peak in X rays. For objectswith a lower-energy synchrotron peak, including flat-spectrum radio quasars, external photon fields coming fromthe accretion disk (among other regions) could be scattered by the electrons resulting in additional radiativecomponents. The potential correlation of these fields with the black-hole and accretion-rate evolution reducesthe value of low-frequency-peaked objects as reliable cosmological gamma-ray beacons (Reimer 2007). Instead,the apparent simplicity of the emission of HSP blazars (with the caveat of their flux variability) makes themthe best reference emitters so far to constrain gamma-ray propagation.
The leading effect on gamma-ray propagation is pair creation on diffuse radiation fields in the Universe. Foran isotropic target field of fixed energy (cid:15) , the cross section is maximum in the comoving frame for a gamma-ray energy E ∼ (2 m e c ) /(cid:15) ∼ × ( (cid:15)/ − (Jauch & Rohrlich 1976). TeV gamma rays thus stronglyinteract with eV, i.e. micron-wavelength photons. This is precisely the wavelength where the cosmic opticalbackground is most intense. The cosmic optical background extends from UV to near infrared wavelengths andis composed of the total emission of stars and galaxies since the end of the cosmic dark ages. Its counterpart atlower energy is the cosmic infrared background, with a peak intensity around ∼ µ m, which comes from thereprocessing of UV-optical light by dust. The cosmic optical and infrared backgrounds are usually referred toas the extragalactic background light (EBL) and constitute the second most intense cosmological backgroundosmic gamma-ray propagation 239after the CMB, which peaks at even lower energy.Measuring the EBL directly proves to be difficult because of the foregrounds from the Galaxy and the solarsystem (zodiacal light) but stringent constraints can be derived accumulating the brightness of known galaxies(e.g., Dole et al. 2006; Madau & Pozzetti 2000; Fazio et al. 2004). Lower limits from galaxy counts typically lieone order of magnitude below upper limits derived from direct measurements. The lower limits are reproducedby EBL models based, e.g., on the local population of galaxies (Franceschini et al. 2008), on the evolution oflarge samples (Dom´ınguez et al. 2011), or on semi-analytical models (Gilmore et al. 2012). Comprehensivesummaries of the observational techniques at these wavelengths as well as of the modeling approaches can befound in Hauser & Dwek (2001), or in Dwek & Krennrich (2013) for a more recent review.To date, the best way to make local foregrounds negligible consists in integrating the optical/near-infraredemission over cosmological distances, as made possible by gamma-ray observations. Indeed, for a density ofphotons n ( (cid:15), z (cid:48) ) in the energy band [ (cid:15) ; (cid:15) + d (cid:15) ] and at a redshift z (cid:48) (i.e. (cid:82) d (cid:15) n ( (cid:15), z (cid:48) ) is a number of photons perunit volume), gamma rays with observed energy E , emitted by a source at a redshift z , are absorbed (throughpair creation) with an optical depth: τ ( E, z ) = (cid:90) z d z (cid:48) d l d z ( z (cid:48) ) (cid:90) + ∞ d (cid:15) n ( (cid:15), z (cid:48) ) (cid:90) − d µ − µ σ ee ( (cid:15), E × (1 + z (cid:48) ) , µ ) (1.1)The first integral represents the distance, with d l/ d z = c/H (1 + z ) (cid:112) Ω Λ + Ω m (1 + z ) in a flat universe.The second is the target density and the third is the cross section with an integration in the center-of-massframe over the angle θ ( µ = cos θ ) between the directions of the target photon and of the gamma ray. Thepair-creation cross section is given by the Bethe-Heitler formula: σ ee ( (cid:15) , (cid:15) , µ ) = 316 σ T (1 − β ) (cid:20) β ( β −
2) + (3 − β ) ln (cid:18) β − β (cid:19)(cid:21) Θ( (cid:15) − (cid:15) th ) (1.2)where Θ is the Heaviside function, (cid:15) th = 2 m e c / (1 − µ ) (cid:15) is the threshold energy and β = 1 − (cid:15) th /(cid:15) . The paircreation on the EBL thus results in an energy and redshift dependent absorption shown in Fig. 2 that can beused to probe the EBL itself (see Sec. 2.2). -ray energy [ TeV ] g -1
10 1 10 r e d s h i ft z -2 -1
101 = 6 t = 3 t = 1 t F
37% x F
5% x F Fig. 2.
Optical depth τ = 1 , ,
40 SF2A 2013
A second order effect on gamma-ray propagation is related to the fate of the pairs created by EBL absorption.For a gamma ray of energy E ∼ γ ∼ , and can upscatter the most intense photon field in its environment, that is theCMB with photons of energy (cid:15) CMB ∼ E (cid:48) ∼ γ (cid:15) CMB ∼ ∗ , an effect often referred to as pair echo.If the intergalactic magnetic field (IGMF) filling the voids is strong enough (typically between 10 − − − G),the pairs should be significantly deflected resulting in an extended emission around the point-like blazar, anemission sometimes referred to as pair halo or beam broadened cascade.The intensity and coherence scale of the IGMF remain largely unknown, though these quantities could be ofcosmological importance (see Durrer & Neronov 2013, for a recent review of the observational and theoreticalprogress). To be amplified through dynamo and compression effects up to their current level, the magneticfields found in galaxies and galaxy clusters would indeed need seeds. These magnetic seeds could arise eitherduring collapses forming the first structures or during the early Universe, be it inflation or electroweak/QCDtransition phases. Current constraints from gamma-ray astronomy are discussed in Sec 2.3.
Exotic physics beyond the standard models of particle physics and cosmology can also be probed within gamma-ray cosmology. Lorentz invariance violation, studied e.g. within quantum gravity frameworks, would result inan alteration of the dispersion relation for VHE gamma rays. The threshold of the pair creation, derived withsimple special relativity arguments, would thus be altered, reducing the EBL opacity to gamma rays (Kifune1999; Jacob & Piran 2008). The absence of such a feature in the spectrum of Mrk 501, as detected by secondgeneration instruments, already set interesting limits (Stecker & Glashow 2001) but the way remains wide openfor refinements with current data.Another crucial test concerns axion-like particles (ALPs). These belong to the class of weakly interactingslim particles (WISPs) that remain a viable alternative to their massive counterparts (WIMPs) as elementaryconstituents of cold dark matter (see e.g. B¨ahre et al. 2013, in the context of the future instrument ALPS-II).Photons and ALPs would indeed mix in the presence of a magnetic field (such as the IGMF or intraclustermagnetic fields), which could modify the apparent absorption of gamma rays (e.g., S´anchez-Conde et al. 2009).A gamma ray converted in an ALP would be unaffected by the EBL absorption and could propagate on largedistances before converting back into a gamma ray and being detected. The universe would then appearmore transparent than expected. Alternatively, for low enough reconversion probabilities, the ALPs would goundetected, potentially increasing the opacity to gamma rays. Recent hints of deviations from a standard EBLabsorption (so-called pair production anomaly) are currently debated in the community, as discussed in Sec. 2.4.
The methodology to constrain EBL with VHE gamma-ray blazars was developed with the flat spectrum radioquasar 3C 279, which would eventually be detected with the sensitivity of third generation instruments. Steckeret al. (1992) performed a straight extrapolation of the HE spectrum of 3C 279 as measured by EGRET upto VHE, arguing that the object should be detectable by the Whipple observatory under the assumption ofno curvature and no EBL absorption. The non detection proved either an intrinsic curvature in the spectrumor/and that EBL absorption, which increases with energy (cf Fig. 2), is strong enough to suppress the flux. Acharacterization of the flux suppression could then provide constraints on the EBL.The detection of Mrk 421 by Whipple enabled the first derivations of upper limits on the EBL (Stecker et al.1993; Dwek & Slavin 1994; Biller et al. 1998). Further studies followed the detection of Mrk 501 by Whipple andHEGRA (Funk et al. 1998; Aharonian et al. 1999), and later on by CAT (Guy et al. 2000; Renault et al. 2001),then complemented by data from the more distant ( z ∼ .
13) HSP blazar H 1426+428 (Dwek & Krennrich ∗ Plasma instabilities resulting in a heating of the medium have nonetheless been invoked as an alternative for the leptons tolose their energy (Broderick et al. 2012). The theoretical debate remains open to date. osmic gamma-ray propagation 2412005). More stringent constraints came from third generation instruments, in particular with the detection of1ES 1101-232 by
H.E.S.S. (Aharonian et al. 2006b) and the combination of spectra from 13 blazars by Mazin& Raue (2007). Absorption at the low energy end of the VHE band has also been constrained by
Fermi -LAT using high redshift blazars and GRBs (Abdo et al. 2010).The combination of HE and VHE measurements led to the most constraining upper limits to date. For z < .
5, the HE band is indeed virtually unaffected by EBL absorption, and
Fermi -LAT data can be used toconstrain the intrinsic emission, while VHE data probe the absorption (as, e.g., in Georganopoulos et al. 2010;Orr et al. 2011; Meyer et al. 2012).
One of the limitations of the studies mentioned above lies in the method originally developed for 3C 279. Sincebroad band EBL absorption and the intrinsic curvature expected in standard emission scenarios both result indownward going observed flux as a function of energy, a characterization of the fall can only result in an upperlimit on the EBL absorption, derived on the basis of a null intrinsic curvature.This limitation has been tackled by the
Fermi -LAT and
H.E.S.S. collaborations with two slightly differentapproaches. In Ackermann et al. (2012), the curvature in the spectra of BL Lac objects was measured in anenergy range where the EBL is ineffective (typically 1 −
50 GeV for 0 . < z < . φ ( E, z ) = φ int ( E ) × exp( − ατ ( E, z, n )) (2.1)where φ is the measured flux, φ int is the extrapolated curved spectrum, the so-called intrinsic spectrum, τ ( E, z, n )is the EBL optical depth for a template EBL model of density n and where α is a normalization factor left freein the fitting procedure.A log-likelihood profile as a function of α is derived for each source, the sum of these profiles resulting inthe combined constraint. Using the model of Franceschini et al. (2008) as a template for the EBL density, adetection (i.e. a normalization factor α differing from zero) at the 5 σ level is performed with 50 sources inthe redshift range 0 . < z < .
6. Because of the limited gamma-ray energy, the EBL effect remains weakfor z < . . < z < . σ measurements shown in gray in Fig. 3, left). One of the strengths of the Fermi -LAT study lies in the investigation of a dozen models, showing that the models of Franceschini et al.(2008), Dom´ınguez et al. (2011) or Gilmore et al. (2012) fit rather well the data, despite of their completelydifferent theoretical grounds. One of the weaknesses could lie in the extrapolation of the low-energy flux, witha fixed intrinsic emission that could result in an overestimation of the significance of the effect † .The H.E.S.S. collaboration measured the EBL absorption with observations of the seven brightest blazars( z < .
2) detected at VHE in the southern hemisphere (H.E.S.S. Collaboration et al. 2013). The analysis is alsobased on likelihood profiles as a function of a normalization factor α . One of the weaknesses of the approachlies in the test of a single EBL model (Franceschini et al. 2008), though retrospectively validated by the studyof Fermi -LAT . Its strength lies in the treatment of the intrinsic spectrum, with the test of five different modelswith free parameters (even allowing for upward going intrinsic emission). The particular behavior of the EBLabsorption between 1 and 10 TeV, due to the depletion of the target EBL photon field between 1 and 10 µ m,constitutes a prominent signature that is detected by a combined likelihood analysis at the 9 σ level. Thenormalization factor of the EBL model as measured by H.E.S.S. and
Fermi -LAT are compared in Fig. 3, left,the top graphic showing the epochs probed for the optical depth. The redshift range 0 . < z < . σ confidence contours shown in Fig. 3, right, where statistical and systematicuncertainties are included. The measurements lie in between lower limits from galaxy counts and upper limitsfrom direct measurements as well as below the region excluded by the gamma-ray analyses discussed in Sec. 2.1.The measurements are dominated by systematic uncertainties below 5 µ m, with an uncertainty of the order of20-30%. At higher wavelengths where statistical uncertainties remain dominant, there is room for improvementusing larger samples of gamma rays at the highest energies. † The limited energy range in which the extrapolation is performed counterweights a bit this caveat.
42 SF2A 2013
Fig. 3. Left:
EBL normalization factor as a function of the distance of the gamma-ray blazars used as beacons. Theblue and gray points represent the
Fermi -LAT measurement and constraints. The red point represents the
H.E.S.S. measurement.
Top graphic adapted from an original from NASA/WMAP Science team . Right:
Spectral energy dis-tribution of the EBL as measured by
H.E.S.S. (red) and
Fermi -LAT (blue) from UV to near infrared wavelengths (1 σ confidence contours). Adapted from
H.E.S.S. Collaboration et al. (2013).
The IGMF is usually characterized by its intensity, B , and its coherence length, λ B . As discussed in Durrer &Neronov (2013), the parameter space remains poorly constrained. A strong IGMF coherent on supra-Mpc scaleswould result in CMB anisotropies extending on large angular scales, resulting in an upper limit of the order B < − G. Large coherence lengths being untestable above the Hubble scale, the parameter space above λ B > Mpc is left out of the search. On theoretical grounds, MHD turbulence tends to dissipate strongmagnetic fields on small scales, heating up the medium, which yields the exclusion region shown in Fig. 4, left.As discussed in Sec. 1.3, joint observations of gamma-ray sources at HE and VHE can probe the intensityof the IGMF. For low enough B , the EBL-absorption induced pairs should barely be deflected, resulting in areprocessing of the signal to HE. The non-observation of this signal, mostly in the spectral energy distributionof the HSP blazar 1ES 0229+200, results in a lower limit of the order of B > − G (Neronov & Vovk2010, shown as a dashed red region in Fig. 4, left), under the assumption of a fixed EBL model and of alongterm steady-state emission of the source. Releasing the latter assumption, Dermer et al. (2011) and Tayloret al. (2011) derived weaker limits of the order 10 − G and 10 − G, the latter being shown as a filled regionin Fig. 4, left. Investigating the underlying hypotheses on the source intrinsic emission, Arlen et al. (2012)concluded than present data on 1ES 0229+200 rule the null IGMF hypothesis out at the 3 σ confidence level.The strong dependence of these constraints on the EBL level is discussed in Vovk et al. (2012), with an increasein the lower limit of almost a factor of 10 for an increase in the EBL level by 30%, which is representative ofthe current uncertainties derived by H.E.S.S. Collaboration et al. (2013) and Ackermann et al. (2012). Some puzzles remain to be solved at VHE. For example, the spectrum of the blazar PKS 1424+240 measured bythe VERITAS collaboration does not seem as absorbed as it should be, given the redshift of the source z ≥ . β and Ly γ absorption in Furniss et al. 2013). Other hints for deviations from nominal EBLopacities have recently been studied in Horns & Meyer (2012) using VHE spectral points. Archival points areosmic gamma-ray propagation 243 ALPS (cid:45)
ISN
Γ(cid:45) burst CASTALPS (cid:45)
IIc
QCD Axion
ALPs from intermed . string scalesALPs as cold dark matter
TeV transparency
WD cooling (cid:45) (cid:45) (cid:45) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) (cid:180) (cid:45) m (cid:64) eV (cid:68) g a Γ (cid:64) G e V (cid:45) (cid:68) Fig. 4. Left:
Excluded regions in parameter space intensity vs coherence length of the IGMF.
Adapted from
Durrer &Neronov (2013).
Right:
Constraints in the parameter space coupling of ALPs with photons vs mass.
Extracted from
B¨ahre et al. (2013). fitted with the model described in Eq. 2.1, assuming a power-law or a log-parabolic intrinsic spectrum and usinga template EBL model with a normalization factor of one. An estimator of the residual to the fit is defined foreach point as: R ( E i ) = F i − φ ( E i ) F i + φ ( E i ) (2.2)where F i is the flux measured at the energy E i . The authors discard an estimator taking into account themeasured uncertainties on the flux, such as ( F i − φ ( E i )) /σ i , because of deviations from the expected normaldistribution of unity width. At high optical depth, or equivalently at high energies for a given redshift, a 4 σ deviation from null residuals is found, indicating an opacity smaller than expected. Systematic uncertainties inthe energy scale decrease the effect to the 3 . σ level, and discarding the last points from each spectrum in thestudy reduces the deviation to the 2 . σ level. In Horns & Meyer (2013), a similar approach is followed usingthree Fermi -LAT events from a gamma-ray burst and two blazars, yielding a 3 . σ effect (preliminary value).The VHE hint has been used by the same group in Meyer et al. (2013) to constrain the parameter space ofALPs, i.e their coupling to photons as a function of the ALP mass, and claim the first lower-limit on the coupling(complex-shaped dark pink region in Fig. 4, right). The region of interest, denoted as “TeV transparency”, liesbelow the regions excluded by the CAST and ALPS-I experiments and is partially excluded by the signal fromSN 1987A. It intersects the light pink band possibly explaining anomalous cooling rates from white dwarves,but remains above theoretical expectations from string theory, cold dark matter scenarios, or standard axionsinvoked as a solution to the absence of CP-violation in QCD, which are in turn probed by the ADMX experimentbetween 2 and 3 µ eV. Not included in this graphic is the recent result from Brun et al. (2013) that excludes alarge fraction of the “TeV transparency” region using VHE data from the HSP blazar PKS 2155-304.As exciting as these hints could be, some systematic uncertainties could affect the estimated significancelevel of the effect. The EBL model is subject to large uncertainties in the energy range concerned (typicallyabove 5 µ m). For the Fermi -LAT event analysis, the method does not seem to account for the energy resolutionof the instruments. For the analysis of VHE spectra, the estimator described in Eq. 2.2 does not accountfor the uncertainties on the flux measurements, which drastically increase with energy. Fitting models topublished spectral points, though common practice, does not properly account for measurement uncertainties44 SF2A 2013and correlations between points ‡ . The last caveat, briefly addressed in the study of systematic uncertainties byHorns & Meyer (2012), concerns the highest energy points in each spectrum. The last point above some minimumsignificance is indeed usually selected when publishing a VHE spectrum, thus biasing the selection towardupward fluctuations with respect to the actual flux. Suppressing these points largely reduces the significance ofthe effect (down to 2 . σ ), potentially suggesting an accumulation of upward going fluctuations. It could alsobe argued that these high-energy points are precisely the ones from which the ALP signal is expected. Alongwith laboratory experiments such as ALPS-II, large improvements in the energy coverage, discussed in Sec. 3,are needed from future generation VHE instruments to decide this issue. From Eq. 1.1, the EBL optical depth first appears as a probe for deviations from the pair-creation cross section(third integral in the equation), as discussed in Sec. 1.4 with Lorentz invariance violation, then as a proxy forthe EBL density (second integral), as discussed in Sec. 2.2 with the measurements of
H.E.S.S. and
Fermi -LAT .Alternatively, it also could in principle be used to constrain distances (first integral).Upper limits on the redshifts of HSP blazars, such as PG 1553+113 (see, e.g., Aharonian et al. 2006a),have been set on a regular basis, going hand in hand with growingly stringent limits on the EBL. Attempts tostandardize the multiwavelength emission of blazars at HE and VHE have been led by Prandini et al. (2010),correlating the HE-VHE spectral differences, which trace EBL absorption, with distance. The authors infer aredshift of z = 0 . ± .
05 for PKS 1424+240 based on the
VERITAS and
Fermi -LAT spectral measurements,emphasizing the puzzle around this object (cf. Sec. 2.4).Constraints on the Hubble parameter H have also been derived using the absorption of gamma-rays fromblazars (Barrau et al. 2008; Dom´ınguez & Prada 2013). Note however that even for an EBL optical depth virtu-ally perfectly measured, the remaining uncertainties on the EBL density (20-30%) would be directly propagatedinto the inferred distances and Hubble parameter. The constraints derived from gamma-ray cosmology are based on a joint modeling of the source emission, EBLabsorption and second order processes, such as the effect of the IGMF or of physics beyond the standard models.Using the spectra derived during multi-wavelength campaigns in the X-ray, HE and VHE gamma-ray bands hasproved particularly constraining for the modeling of blazars’ emission. Using this multi-wavelength synergy, inparticular the evolution of HE-VHE spectra with redshift, Sanchez et al. (2013) have recently confirmed the
H.E.S.S. and
Fermi -LAT measurements.Using in addition X-ray observations, Mankuzhiyil et al. (2010) and Dom´ınguez et al. (2013) modeled thebroad band spectra of blazars within synchrotron self-Compton scenarios and jointly constrained the EBLoptical depth, though the same caveat as in Sec. 2.4 (uncertainties and correlations between the points) couldbe raised. The multi-wavelength approach is crucial for probes of the IGMF based on pair-echo, with thereprocessing of the VHE signal to HE. The IGMF intensity could nonetheless potentially be constrained up to10 − G solely with CTA data, using pair-haloes / beam-broadened-cascades approaches (Durrer & Neronov2013; Sol et al. 2013b).
The lowering of the energy threshold of ground-based gamma-ray instruments below 100 GeV is a crucialstep in widening the sample of probes of the Universe opacity to gamma rays. To increase the sensitivity toCherenkov light from low-energy gamma-ray induced atmospheric showers, the VERITAS collaboration optedfor high quantum efficiency photomultipliers, while the MAGIC and H.E.S.S. collaborations opted for large sizetelescopes (28 m diameter for the recently built CT5 telescope of the H.E.S.S. II array). A low energy thresholdimplies a larger gamma-ray horizon, as shown e.g. in Fig. 2. VHE instruments could then probe the EBL ‡ VHE points are generally highly correlated due to the steepness of the spectra and the limited energy resolution that implyleakage of events from low to higher energies. Spectral points derived with a maximum likelihood analysis should be considered asresiduals to the fit. Therefore, fitting a model to these residuals may not be the most statistically sound approach. osmic gamma-ray propagation 245optical depth using blazars at 0 . < z < .
5, which remains uncharted territory (cf. Fig. 3, left). Measuringgamma-ray opacity in various redshift bands would result in a tomography of the EBL density, enabling thecomparison of the EBL evolution with e.g. the star formation history (Raue & Meyer 2012).
Based on the
Fermi -LAT population of extragalactic sources, the population detected by ground-based instru-ments should become less and less dominated by HSP objects, allowing tests of unification schemes of blazars.These investigations, which also constitute a fundamental science case for CTA (see Reimer & B¨ottcher 2013;Sol et al. 2013a, and references therein), will deepen our knowledge on potential beacons and thus directlybenefit gamma-ray cosmology. The knowledge of the Universe opacity at VHE will also directly impact thephysics of gamma-ray bursts (see Inoue et al. 2013; M´esz´aros 2013, in the context of CTA), which should bedetected at a rate of a few per year by CTA (Gilmore et al. 2013).CTA, with its large-size telescopes enabling an energy threshold of 30 GeV or less, its medium-size telescopesboosting the sensitivity by a factor of 10 in the core energy range, and small size telescopes paving a fewkilometer-square area that will probe energies above 10 TeV, will take gamma-ray cosmology to its full maturity(see Mazin et al. 2013; Ellis & Mavromatos 2013, for discussions on EBL and Lorentz invariance in the CTAera). CTA will in particular probe the EBL density above 5 µ m, where diffuse emission from polycyclic aromatichydrocarbons could be detected and where physics beyond the standard models could be unveiled, yieldingobservables for quantum gravity theories or clues on the elusive question of the nature of dark matter. JB would like to thank D. Williams for his useful suggestions, which improved this text.
References
Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010, ApJ, 723, 1082Ackermann, M., Ajello, M., Allafort, A., et al. 2012, Science, 338, 1190Aharonian, F., Akhperjanian, A. G., Bazer-Bachi, A. R., et al. 2006a, A&A, 448, L19Aharonian, F., Akhperjanian, A. G., Bazer-Bachi, A. R., et al. 2006b, Nature, 440, 1018Aharonian, F. A., Akhperjanian, A. G., Barrio, J. A., et al. 1999, A&A, 349, 11Antonucci, R. 1993, ARA&A, 31, 473Arlen, T. C., Vassiliev, V. V., Weisgarber, T., Wakely, S. P., & Yusef Shafi, S. 2012, arXiv:1210.2802B¨ahre, R., D¨obrich, B., Dreyling-Eschweiler, J., et al. 2013, Journal of Instrumentation, 8, 9001Band, D. L. & Grindlay, J. E. 1985, ApJ, 298, 128Barrau, A., Gorecki, A., & Grain, J. 2008, MNRAS, 389, 919Biller, S. D., Buckley, J., Burdett, A., et al. 1998, Physical Review Letters, 80, 2992Broderick, A. E., Chang, P., & Pfrommer, C. 2012, ApJ, 752, 22Brun, P., Wouters, D., & for the H. E. S. S. collaboration. 2013, arXiv:1307.6068Dermer, C. D., Cavadini, M., Razzaque, S., et al. 2011, ApJ, 733, L21Dole, H., Lagache, G., Puget, J.-L., et al. 2006, A&A, 451, 417Dom´ınguez, A., Finke, J. D., Prada, F., et al. 2013, ApJ, 770, 77Dom´ınguez, A. & Prada, F. 2013, ApJ, 771, L34Dom´ınguez, A., Primack, J. R., Rosario, D. J., et al. 2011, MNRAS, 410, 2556Durrer, R. & Neronov, A. 2013, A&A Rev., 21, 62Dwek, E. & Krennrich, F. 2005, ApJ, 618, 657Dwek, E. & Krennrich, F. 2013, Astroparticle Physics, 43, 112Dwek, E. & Slavin, J. 1994, ApJ, 436, 696Ellis, J. & Mavromatos, N. E. 2013, Astroparticle Physics, 43, 50Fazio, G. G., Ashby, M. L. N., Barmby, P., et al. 2004, ApJS, 154, 39Franceschini, A., Rodighiero, G., & Vaccari, M. 2008, A&A, 487, 837Funk, B., Magnussen, N., Meyer, H., et al. 1998, Astroparticle Physics, 9, 97Furniss, A., Williams, D. A., Danforth, C., et al. 2013, ApJ, 768, L31Georganopoulos, M., Finke, J. D., & Reyes, L. C. 2010, ApJ, 714, L157Ghisellini, G. 2013, arXiv:1309.4772
46 SF2A 201346 SF2A 2013