The Extragalactic Background Light · The Extragalactic Background Light. Elisa Pueschel ICRC 2017, Busan, South Korea 19.07.2017 Constraints from TeV Blazar Observations H. Dole

The Extragalactic Background Light.

Elisa Pueschel ICRC 2017, Busan, South Korea 19.07.2017

Constraints from TeV Blazar ObservationsH. Dole et al.: The Cosmic Infrared Background Resolved by Spitzer. 13

• A stacking analysis in three fields covering 0.85 squaredegrees including a sample of 19181 MIPS 24 µm sourceswith S24 ≥ 60 µJy lets us probe faint 70 and 160 µm galax-ies one order of magnitude below the confusion level andwith a high signal-to-noise ratio. We take into account inour noise budget uncertainties coming from: photometry,calibration systematics, and large-scale structure.

• 24 µm galaxies down to S24 = 60 µJy contribute 79%,92%, 69% of the CIB at respectively 24, 70 and 160 µm(using 2.7, 6.4 and 15.4 nW m−2 sr−1 as the total CIBvalues at 24, 70 and 160 µm, respectively). This is the firstdirect measurement of the contribution of MIR-selectedgalaxies to the FIR background.

• We derive the contributions to the CIB by flux densitybin, and show good agreement between our stacking anal-ysis and the published source counts. This is a strong con-straint for models. Moreover, we show that the CIB will bemainly resolved at flux densities of about S70 ∼ 0.9 mJyand S160 ∼ 3 mJy at 70 and 160 µm, respectively.

• We directly measure that the total CIB, peaking near150 µm, is largely resolved into MIR galaxies. Otherworks (Perez-Gonzalez et al., 2005; Le Floc’h et al., 2005;Caputi et al., 2006, especially) show that these MIPS24 µm sources are ∼ 3 × 1011 L⊙ LIRGs distributed atredshifts z ∼ 1, with stellar masses of about 3 × 1010 to3× 1011 M⊙ and specific star formation rates in the range0.1 to 1 Gyr−1.

• Using constant color ratios 160/24 and 70/24 for MIRgalaxies fainter than 60 µJy, we derive new conservativelower limits to the CIB at 70 and 160 µm including thefaint IR galaxies undetected at 24 µm: 7.1±1.0 and 13.4±1.7 nW m−2 sr−1, respectively. These new estimates agreewithin 13% with the Lagache et al. (2004) model.

• Using these new estimates for the 70 and 160µm CIB,we show that our stacking analysis down to S24 ≥ 60 µJyresolves >75% of the 70 and 160 µm CIB.

• Upper limits from high-energy experiments and directdetections together with lower limits from galaxy countsand stacking analysis give strong constraints on the EBLSED.

• We estimate the Extragalactic Background Light(EBL) Spectral Energy Distribution (SED) permittedzone (between lower and upper limits), and measurethe optical background (COB) to be in the range 19.5-35.5 nW m−2 sr−1, and the IR background (CIB) in therange 24 to 27.5 nW m−2 sr−1. The ratio COB/CIB thuslies between 0.7 and 1.5.

• We integrate our best estimate of the COB and the CIB,and obtain respectively 23 and 24 nW m−2 sr−1; We finda COB/CIB ratio close to unity.

• The galaxy formation and evolution processes have pro-duced photons equivalent in brightness to 5% of the CMB,with equal amounts from direct starlight (COB) and fromdust-reprocessed starlight (CIB). We compute that theEBL produces on average 115 infrared photons per visiblephoton.

10-1 100 101 102 103 104 105

Wavelength λ [µm]

10-10

10-9

10-8

10-7

10-6

W m

-2 s

r-1

106 105 104 103 102 101Frequency ν [GHz]

10-1 100 101 102 103 104 105

Wavelength λ [µm]

10-10

10-9

10-8

10-7

10-6

W m

-2 s

r-1

COB CIB

CMB

23 24

960

Fig. 14. Schematic Spectral Energy Distributions of themost important (by intensity) backgrounds in the uni-verse, and their approximate brightness in nW m−2 sr−1

written in the boxes. From right to left: the CosmicMicrowave Background (CMB), the Cosmic InfraredBackground (CIB) and the Cosmic Optical Background(COB).

Acknowledgements. This work is based on observations madewith the Spitzer Observatory, which is operated by the JetPropulsion Laboratory, California Institute of Technology un-der NASA contract 1407. We thank the funding from theMIPS project, which is supported by NASA through theJet Propulsion Laboratory, subcontract #1255094. This workalso benefited from funding from the CNES (Centre Nationald’Etudes Spatiales) and the PNC (Programme Nationalde Cosmologie). We thank Jim Cadien for the great helpin the data processing. We thank Nabila Aghanim, HerveAussel, Noel Coron, Daniel Eisenstein, David Elbaz, CharlesEngelbracht, Dave Frayer, Karl Gordon, Nicolas Ponthieu,Martin Schroedter, and Xianzhong Zheng for fruitful discus-sions.

References

Aharonian, F, Akhperjanian, A. G, Bazer-Bachi, A. R,Beilicke, M, & H.E.S.S Collaboration. Nature, 2006,accepted, astro-ph/0508073.

Bernstein, R. A, Freedman, W. L, & Madore, B. F. 2002,ApJ, 571:107.

Brown, T. M, Kimble, R. A, Ferguson, H. C, Gardner,J. P, Collins, N. R, & Hill, R. S. 2000, AJ, 120:1153.

Cambresy, L, Reach, W. T, Beichman, C. A, & Jarrett,T. H. 2001, ApJ, 555:563.

Caputi, K. I, Dole, H, Lagache, G, et al., 2006, ApJ,637:727.

Chary, R & Elbaz, D. 2001, ApJ, 556:562.Chary, R, Casertano, S, Dickinson, M. E, et al., 2004,

Astrophysical Journal Supplement Series, 154:80.Dole, H, Lagache, G, & Puget, J. L. 2003, ApJ, 585:617.Dole, H, Le Floc’h, E, Perez-Gonzalez, P. G, et al., 2004a,

Astrophysical Journal Supplement Series, 154:87.

Dole 2006

Raue 2011

19.07.2017ICRC 2017Elisa Pueschel

Extragalactic Background Light

2

Imprint from reionization, star formation, galaxy evolution, emission by active galactic nuclei

Unresolved sources? Dark matter decay? Exotic physics? Connected to other diffuse fields (X-ray, radio, neutrino)

Credit: NASA


Extragalactic Background Light

3

> COB = Cosmic optical background > Light from stars, galaxies, etc

> CIB = Cosmic infrared background > Light reprocessed by dust

H. Dole et al.: The Cosmic Infrared Background Resolved by Spitzer. 13

• A stacking analysis in three fields covering 0.85 squaredegrees including a sample of 19181 MIPS 24 µm sourceswith S24 ≥ 60 µJy lets us probe faint 70 and 160 µm galax-ies one order of magnitude below the confusion level andwith a high signal-to-noise ratio. We take into account inour noise budget uncertainties coming from: photometry,calibration systematics, and large-scale structure.

• 24 µm galaxies down to S24 = 60 µJy contribute 79%,92%, 69% of the CIB at respectively 24, 70 and 160 µm(using 2.7, 6.4 and 15.4 nW m−2 sr−1 as the total CIBvalues at 24, 70 and 160 µm, respectively). This is the firstdirect measurement of the contribution of MIR-selectedgalaxies to the FIR background.

• We derive the contributions to the CIB by flux densitybin, and show good agreement between our stacking anal-ysis and the published source counts. This is a strong con-straint for models. Moreover, we show that the CIB will bemainly resolved at flux densities of about S70 ∼ 0.9 mJyand S160 ∼ 3 mJy at 70 and 160 µm, respectively.

• We directly measure that the total CIB, peaking near150 µm, is largely resolved into MIR galaxies. Otherworks (Perez-Gonzalez et al., 2005; Le Floc’h et al., 2005;Caputi et al., 2006, especially) show that these MIPS24 µm sources are ∼ 3 × 1011 L⊙ LIRGs distributed atredshifts z ∼ 1, with stellar masses of about 3 × 1010 to3× 1011 M⊙ and specific star formation rates in the range0.1 to 1 Gyr−1.

• Using constant color ratios 160/24 and 70/24 for MIRgalaxies fainter than 60 µJy, we derive new conservativelower limits to the CIB at 70 and 160 µm including thefaint IR galaxies undetected at 24 µm: 7.1±1.0 and 13.4±1.7 nW m−2 sr−1, respectively. These new estimates agreewithin 13% with the Lagache et al. (2004) model.

• Using these new estimates for the 70 and 160µm CIB,we show that our stacking analysis down to S24 ≥ 60 µJyresolves >75% of the 70 and 160 µm CIB.

• Upper limits from high-energy experiments and directdetections together with lower limits from galaxy countsand stacking analysis give strong constraints on the EBLSED.

• We estimate the Extragalactic Background Light(EBL) Spectral Energy Distribution (SED) permittedzone (between lower and upper limits), and measurethe optical background (COB) to be in the range 19.5-35.5 nW m−2 sr−1, and the IR background (CIB) in therange 24 to 27.5 nW m−2 sr−1. The ratio COB/CIB thuslies between 0.7 and 1.5.

• We integrate our best estimate of the COB and the CIB,and obtain respectively 23 and 24 nW m−2 sr−1; We finda COB/CIB ratio close to unity.

• The galaxy formation and evolution processes have pro-duced photons equivalent in brightness to 5% of the CMB,with equal amounts from direct starlight (COB) and fromdust-reprocessed starlight (CIB). We compute that theEBL produces on average 115 infrared photons per visiblephoton.

10-1 100 101 102 103 104 105

Wavelength λ [µm]

10-10

10-9

10-8

10-7

10-6

W m

-2 s

r-1

106 105 104 103 102 101Frequency ν [GHz]

10-1 100 101 102 103 104 105

Wavelength λ [µm]

10-10

10-9

10-8

10-7

10-6

W m

-2 s

r-1

COB CIB

CMB

23 24

960

Fig. 14. Schematic Spectral Energy Distributions of themost important (by intensity) backgrounds in the uni-verse, and their approximate brightness in nW m−2 sr−1

written in the boxes. From right to left: the CosmicMicrowave Background (CMB), the Cosmic InfraredBackground (CIB) and the Cosmic Optical Background(COB).

Acknowledgements. This work is based on observations madewith the Spitzer Observatory, which is operated by the JetPropulsion Laboratory, California Institute of Technology un-der NASA contract 1407. We thank the funding from theMIPS project, which is supported by NASA through theJet Propulsion Laboratory, subcontract #1255094. This workalso benefited from funding from the CNES (Centre Nationald’Etudes Spatiales) and the PNC (Programme Nationalde Cosmologie). We thank Jim Cadien for the great helpin the data processing. We thank Nabila Aghanim, HerveAussel, Noel Coron, Daniel Eisenstein, David Elbaz, CharlesEngelbracht, Dave Frayer, Karl Gordon, Nicolas Ponthieu,Martin Schroedter, and Xianzhong Zheng for fruitful discus-sions.

References

Aharonian, F, Akhperjanian, A. G, Bazer-Bachi, A. R,Beilicke, M, & H.E.S.S Collaboration. Nature, 2006,accepted, astro-ph/0508073.

Bernstein, R. A, Freedman, W. L, & Madore, B. F. 2002,ApJ, 571:107.

Brown, T. M, Kimble, R. A, Ferguson, H. C, Gardner,J. P, Collins, N. R, & Hill, R. S. 2000, AJ, 120:1153.

Cambresy, L, Reach, W. T, Beichman, C. A, & Jarrett,T. H. 2001, ApJ, 555:563.

Caputi, K. I, Dole, H, Lagache, G, et al., 2006, ApJ,637:727.

Chary, R & Elbaz, D. 2001, ApJ, 556:562.Chary, R, Casertano, S, Dickinson, M. E, et al., 2004,

Astrophysical Journal Supplement Series, 154:80.Dole, H, Lagache, G, & Puget, J. L. 2003, ApJ, 585:617.Dole, H, Le Floc’h, E, Perez-Gonzalez, P. G, et al., 2004a,

Astrophysical Journal Supplement Series, 154:87.

Dole 2006

Credit: A. Jaffe

Refers to one part of the diffuse photon background

EBL density evolves with time/redshift


Modeling the Extragalactic Background Light

4

5.3. Cosmic chemical evolution (CCE) models

CCE models treat the universe as a closed system in which allgalaxies within a large comoving volume element are representedby their basic ingredients: stars, interstellar gas, metallicity, andradiation. Chemical evolution equations, analogous to those usedto follow the chemical evolution of the Galaxy (e.g.[46,231,193]),are used to follow the evolution of the average stellar, gaseous,and radiative contents in each comoving volume in a self consis-tent manner. CCE models were pioneered by Pei and Fall [195],and most recently updated by Pei et al. [196]. Inputs parametersfor their model are the mean rest frame UV luminosity density asa function of redshift, and the mass of the ISM gas as determinedfrom H I column densities derived from studies of quasar absorp-tion lines through damped Lya systems. The decrease in the ISMgas with redshift and the UV luminosity density were used to de-rive a solution for the evolution of the CSFR with redshift whichis consistent with that determined from the extinction-correctedHa, and with SCUBA 850 and ISO 15 lm surveys. Similar to FE mod-els, population synthesis models were then used to calculate thestellar SED at each redshift, and an LMC extinction law wasadopted to calculate the fraction of starlight absorbed by the dust.A power-law distribution in dust temperature was used to calcu-late the spectrum of the reradiated IR emission. The model repro-duced various observational constraints, including the comovingrest-frame 0.44, 1.0, and 2.2 lm spectral luminosity densities inthe !0–2 redshift interval, the 12, 25, 60, and 100 lm local lumi-nosity densities; and the mean abundance of metals in dampedLya systems in the !0.4–3.5 redshift interval.

5.4. Semi-analytical (SA) models

SA models follow the formation and evolution of galaxies in acold dark matter Lambda dominated (KCDM) universe using thecosmological parameters derived from the 5-year Wilkinson Micro-wave Anisotropy Probe (WMAP5) observations [128] as the initialconditions. SA models then follow the growth and merger of darkmatter halos, and the emergence of galaxies which form as bary-onic matter falls into the potential wells of these halos. The fateof the infalling gas is determined by many different processes:the formation of stars in a multiphase interstellar medium, AGNand supernovae feedback processes that quench their formation,the evolution of the stellar radiation field, the heating and coolingof the interstellar medium and its chemical enrichment, the ex-change of material with the intergalactic medium through infalland galactic winds, and the growth of the central black hole. Adescription of recent developments and references to previouswork can be found in [222]. Model prediction are compared to abasic set of observational constraints such as the observed charac-teristics of galaxies: their morphology, colors, and spectral energydistribution, and morphology; and their integrated cosmologicalproperties: their number counts and luminosity function in differ-ent wavebands and redshifts, their mass function, the cosmic starformation rate, and the EBL generated by them. As in all EBL mod-els, determination of the galaxies’ SED is complicated by the de-tailed microscopic and large scale parameters needed to calculatethe amount of starlight that is absorbed by dust, and the spectrumof the reradiated emission. Recent SA models have combined themodels for galaxy formation with radiative transfer models todetermine the galaxies’ SED [103,102,222,241].

SA models are inherently complex, incorporating a large num-ber of physical processes, some poorly known, to derive galaxyproperties. However, they are the most physically motivated mod-els, and quite successful in reproducing a large number of observa-tional constraints.

5.5. Comparison of model predictions with observations

A detailed comparison of all model types with EBL limits andobservations was presented by Hauser and Dwek [124]. Here wewill represent mostly the models that have been developed sincethen: BE models by Stecker and Scully [229], Franceschini et al.[107], and Domínguez et al. [81]; The FE model of Finke et al.[99]; and the SA model of Gilmore et al. [117]. Fig. 9 comparesthe various models to the current limits and observations of theEBL. In general, all models, except for the BE models of Steckeret al. provide adequate fits to the EBL.

6. EBL constraints from c-ray observations of blazars

The attenuation of c-rays by the EBL can in principle be used todetermine the EBL intensity at wavelengths corresponding to thec-ray observations. Neglecting the possible scattering or produc-tion of second generation c-ray photons along the line of sight tothe blazar, the intrinsic c-ray flux from the blazar, FIntðEcÞ, can berelated to the observed one, FobsðEcÞ by:

FobsðEcÞ ¼ FIntðEcÞ exp½&sccðEcÞ' ð15Þ

where the optical depth, scc, is given by Eq. (9). Determination ofthe EBL assumes that all the attenuation is caused by interactionwith the EBL, instead of photons in or around the vicinity of the bla-zar. Furthermore, it requires knowledge of the intrinsic blazar spec-trum. Assuming that all the attenuation is attributed to the EBL,several upper limits have been derived on the EBL intensity by mak-ing various assumptions on the intrinsic blazar spectrum. The c-rayderived EBL limits are compared to those derived from UV to submillimeter observations in Fig. 10, and described below.

Fixed power law: Early observations of the blazars Mrk 421 and3C 279 suggested that their !GeV–TeV spectrum could be approx-imated by a single power law ([201,227], respectively). If so, thenany deviations of the observations from the extrapolated powerlaw to higher energies should be attributed to EBL attenuation.Stecker and de Jager [226] derived an upper limits at 1–5 lm of10 nW m&2 sr&1, assuming that a straight power law extrapolationof the spectrum of Mrk 421 from GeV energies obtained from theEGRET [163] with an index of C ¼ 1:96( 0:14 holds up to theTeV regime, where the index was measured by the Whipple Collab-oration to be C ¼ 2:25( 0:19. Biller et al. [61] included the statis-tical uncertainties of the GeV spectrum, and demonstrated vastlydifferent extrapolations with significantly higher upper limits,thereby yielding conservative upper limit to the EBL in themid-IR at 10 lm.

Fig. 9. Models of the EBL are compared to observational limits on the EBL.

124 E. Dwek, F. Krennrich / Astroparticle Physics 43 (2013) 112–133

Dwek & Krennrich 2013

> Backward evolution models > Start with local luminosity

density + galaxy number counts

> e.g. Stecker et al., Franceschini et al., Domínguez et al.

Multiple approaches to predict the EBL intensity & spectrum

> Forward evolution models > Start with cosmic star

formation rate, radiative transfer + population synthesis models

> e.g. Finke et al.

> Semi-analytical models > ΛCDM universe > e.g. Gilmore et al.


Measuring the Extragalactic Background Light

5

m]µ [λ1−10 1 10 210

]-1 s

r-2

) [nW

mλ(

ν Iν

1

10

Lower limitsUpper limits

> Direct measurements of night sky background > Sensitive to foreground

contamination > Upper limits > COBE/FIRAS/DIRBE

>Galaxy counts in magnitude bands > Stacking analysis > Fluctuation analysis

>Miss diffuse/unknown components

> Lower limits


Indirect Measurements of EBL with Gamma-ray Emitters

6

Raue 2011

Background Energy [eV]-410 -310 -210 -110 1 10

]-1

pc

-1) [

eVŒ( gg

Inte

gran

d Q

-1110

-1010

-910

-810

-710

-610

-510

-410

-310 ]-3

dn/

dE [e

V c

m2

Bac

kgro

und

Spe

ctru

m E

-310

-210

-110

100 TeV10 TeV1 TeV100 GeVBackgrounds

λeγ

1

λeγ=

! ∞

0dϵ

nϵ(ϵ)m4c8

8βE2ϵ2

! x+

x−dxxσeγ(x) ≡

! ∞

0dϵQeγ(ϵ),

x± = 2Eϵ(1 ± β)/m2c4 cos θ = ±1

Qeγ(ϵ)

> Photons from distant sources interact with EBL photons > γγEBL → e+e-

> Consider mean free path for e+/e- pair production > Probe CIB and COB with primary

photons 100 GeV - 100 TeV

1

��=

Z 1

0d✏ Q��(✏)

Credit: T. Weisgarber

CIB COB


Blazars: A Convenient Probe

7

sccðEc; zÞ ¼ $ logFðEc; zÞobs

FðEcÞint

! "ð16Þ

The opacity can be determined from models of the EBL if its evolu-tion with redshift is known and, independently, from c-ray observa-tions if the intrinsic blazar spectrum is known. Concordancebetween these two independent determinations of scc can serveas a test for the validity of the underlying assumptions in eachmethod.

Figs. 12 and 13 depict the evolution of the comoving intensity ofthe EBL, the corresponding evolution of the proper number densityof background photons, the optical depth to blazars at various red-shift, and the corresponding attenuation factor. Results are plottedfor the BE evolution model of Franceschini et al. [107] and the BEevolution model of Domínguez et al. [81].

Determining the c-ray opacity from observations requiresknowledge of the intrinsic blazar spectrum. Differences betweenthe observed and expected flux at a given energy Ec would thenbe simply attributed to EBL attenuation. Figs. 12 and 13 show thatthe sharp drop of the EBL intensity at UV and shorter wavelengthsrenders the universe almost transparent to GeV photons. Conse-quently, the observed %1–50 GeV spectrum is very likely theintrinsic blazar spectrum. So instead of assuming a theoretical limiton the spectral index, one can use the GeV – 10s of GeV energyspectral slope from Fermi data as a proxy for the intrinsic spectraat TeV energies.

Assuming that this power law can be extrapolated from GeV toTeV energies, one can derive the TeV optical depth to the observedblazar. This approach was used by Georganopoulos et al. [116] andin method 1 in [192] to set firm upper limits on EBL models usingthe GeV to TeV spectra of PKS 2155-304 (z ¼ 0:116) and 1ES1218+304 (z ¼ 0:182). Assuming that the GeV spectrum is unatten-uated by the EBL, [169] used optical, X-ray and GeV data to modelthe TeV flux of PKS 2155-304 using a one-zone SSC model. Com-parison of the model results with observations, they derived the

TeV opacity to this blazar, and found it to be consistent with mostEBL models.

Fig. 14 compares the dependence of the optical depth derivedfrom EBL models (hatched curves) to that derived for select bla-zars: Mrk 501, 1ES 1218+304, and 3C 66A. Each hatched bandspans the range of optical depths predicted by the EBL models ofFranceschini et al. [107], Finke et al. [99], Domínguez et al. [81],and Gilmore et al. [117]. The colored dots represent the opticaldepths derived from the c-ray observations of the three blazars.The intrinsic blazar spectrum was assumed to be a power lawdetermined by the observed flux at 1 GeV and the spectral index,CGeV . The observed flux in the TeV range was assumed to be apower law with a spectral index CTeV (see Table 2). The c-ray opac-ity in the TeV range was then derived from Eq. (11). The band ofopacities for each blazar was obtained by performing 100 MonteCarlo simulations of the intrinsic and observed spectra using theuncertainties in the spectral indices and c-ray energies intoaccount.

The figure shows that the c-ray derived optical depths of Mrk501 and 1ES 1218-304 are in general agreement with model pre-diction. The discrepancy between the EBL and the c-ray derivedoptical depth for 3C 66A is typical of most blazars listed in Table 2.We note that the redshift to 3C 66A is still somewhat uncertain [9].The convergence between observational limits on the EBL andmodels suggests that the origin of the discrepancy can be mostlyattributed to our still incomplete knowledge of the intrinsic spec-tra of blazars.

The EBL not only affects the c-ray spectra of individual c-raysources, but also the spectrum of the extragalactic c-ray back-ground (EGRB) which consists of the cumulative contribution of re-solved and unresolved sources and a possible truly diffuseemission component.

Recently, the Fermi Large Area Telescope (LAT) provided a newmeasurement of the diffuse c-ray background (DGB) at energiesbetween 0.2 and 100 GeV [7], obtained by the subtraction of

Fig. 12. Basic EBL model results by Franceschini et al. [107]: Top left: the comoving EBL and CMB intensities versus wavelength for different redshifts. Top right: the propernumber density of EBL and CMB photons versus energy for the same grid of redshifts as the previous panel. Bottom left: the c-ray opacity versus energy, Ec for differentredshifts. Bottom right: the amount of attenuation versus energy for the same grid of redshifts as the previous panel. The figure illustrates the change in the slope of scc atenergies corresponding to the wavelength at which the slope of the EBL spectrum changes.

E. Dwek, F. Krennrich / Astroparticle Physics 43 (2013) 112–133 127

sccðEc; zÞ ¼ $ logFðEc; zÞobs

FðEcÞint

! "ð16Þ

The opacity can be determined from models of the EBL if its evolu-tion with redshift is known and, independently, from c-ray observa-tions if the intrinsic blazar spectrum is known. Concordancebetween these two independent determinations of scc can serveas a test for the validity of the underlying assumptions in eachmethod.

Figs. 12 and 13 depict the evolution of the comoving intensity ofthe EBL, the corresponding evolution of the proper number densityof background photons, the optical depth to blazars at various red-shift, and the corresponding attenuation factor. Results are plottedfor the BE evolution model of Franceschini et al. [107] and the BEevolution model of Domínguez et al. [81].

Determining the c-ray opacity from observations requiresknowledge of the intrinsic blazar spectrum. Differences betweenthe observed and expected flux at a given energy Ec would thenbe simply attributed to EBL attenuation. Figs. 12 and 13 show thatthe sharp drop of the EBL intensity at UV and shorter wavelengthsrenders the universe almost transparent to GeV photons. Conse-quently, the observed %1–50 GeV spectrum is very likely theintrinsic blazar spectrum. So instead of assuming a theoretical limiton the spectral index, one can use the GeV – 10s of GeV energyspectral slope from Fermi data as a proxy for the intrinsic spectraat TeV energies.

Assuming that this power law can be extrapolated from GeV toTeV energies, one can derive the TeV optical depth to the observedblazar. This approach was used by Georganopoulos et al. [116] andin method 1 in [192] to set firm upper limits on EBL models usingthe GeV to TeV spectra of PKS 2155-304 (z ¼ 0:116) and 1ES1218+304 (z ¼ 0:182). Assuming that the GeV spectrum is unatten-uated by the EBL, [169] used optical, X-ray and GeV data to modelthe TeV flux of PKS 2155-304 using a one-zone SSC model. Com-parison of the model results with observations, they derived the

TeV opacity to this blazar, and found it to be consistent with mostEBL models.

Fig. 14 compares the dependence of the optical depth derivedfrom EBL models (hatched curves) to that derived for select bla-zars: Mrk 501, 1ES 1218+304, and 3C 66A. Each hatched bandspans the range of optical depths predicted by the EBL models ofFranceschini et al. [107], Finke et al. [99], Domínguez et al. [81],and Gilmore et al. [117]. The colored dots represent the opticaldepths derived from the c-ray observations of the three blazars.The intrinsic blazar spectrum was assumed to be a power lawdetermined by the observed flux at 1 GeV and the spectral index,CGeV . The observed flux in the TeV range was assumed to be apower law with a spectral index CTeV (see Table 2). The c-ray opac-ity in the TeV range was then derived from Eq. (11). The band ofopacities for each blazar was obtained by performing 100 MonteCarlo simulations of the intrinsic and observed spectra using theuncertainties in the spectral indices and c-ray energies intoaccount.

The figure shows that the c-ray derived optical depths of Mrk501 and 1ES 1218-304 are in general agreement with model pre-diction. The discrepancy between the EBL and the c-ray derivedoptical depth for 3C 66A is typical of most blazars listed in Table 2.We note that the redshift to 3C 66A is still somewhat uncertain [9].The convergence between observational limits on the EBL andmodels suggests that the origin of the discrepancy can be mostlyattributed to our still incomplete knowledge of the intrinsic spec-tra of blazars.

The EBL not only affects the c-ray spectra of individual c-raysources, but also the spectrum of the extragalactic c-ray back-ground (EGRB) which consists of the cumulative contribution of re-solved and unresolved sources and a possible truly diffuseemission component.

Recently, the Fermi Large Area Telescope (LAT) provided a newmeasurement of the diffuse c-ray background (DGB) at energiesbetween 0.2 and 100 GeV [7], obtained by the subtraction of

Fig. 12. Basic EBL model results by Franceschini et al. [107]: Top left: the comoving EBL and CMB intensities versus wavelength for different redshifts. Top right: the propernumber density of EBL and CMB photons versus energy for the same grid of redshifts as the previous panel. Bottom left: the c-ray opacity versus energy, Ec for differentredshifts. Bottom right: the amount of attenuation versus energy for the same grid of redshifts as the previous panel. The figure illustrates the change in the slope of scc atenergies corresponding to the wavelength at which the slope of the EBL spectrum changes.

E. Dwek, F. Krennrich / Astroparticle Physics 43 (2013) 112–133 127

> Blazars detected to cosmological distances > z=3 at high energies (<100 GeV) > z~1 at very high energies (>100 GeV)

>Optical depth increases with energy and redshift > Depends on γγ interaction cross-section and number density of EBL photons

– 58 –

108 1011 1014 1017 1020 1023 1026 1029

ν [Hz]

10-14

10-13

10-12

10-11

10-10

10-9

10-8

νFν [

erg

s-1 c

m-2

]

Fig. 11.— SED of Mrk 421 with two 1-zone SSC model fits obtained with different minimumvariability timescales: tvar = 1 day (red curve) and tvar = 1 hour (green curve) . The

parameter values are reported in Table 4. See text for further details.

Dwek & Krennrich 2013

Credit: NASA

Abdo 2011

Flux & spectral variability


Extragalactic Light Imprint on Blazar Spectra

8

Intrinsic spectrum: Observed spectrum:

dN

dE/

✓dN

dE int

◆exp(�⌧��)

dN

dE/ E��,

dN

dE/ E��

exp

⇣� E

EC

⌘, ...?

Use blazar observations to learn about EBL

Need to understand EBL to understand blazar spectral properties!

E

dN/d

E

E

dN/d

EJoint fit of EBL parameters & spectral properties or

Minimal assumptions on intrinsic spectral properties + multiple sources 1. No convex spectral shapes 2. Extrapolate Fermi-LAT spectra to > 100 GeV


Two Common Approaches

9

> “Model-dependent” > Derive opacity from EBL shape > Allow scale factor to float

> “Model-independent” > Assume an EBL shape at z=0 > Calculate opacity at relevant

redshifts > Alternative - wavelength-resolved

analysis

5.3. Cosmic chemical evolution (CCE) models

CCE models treat the universe as a closed system in which allgalaxies within a large comoving volume element are representedby their basic ingredients: stars, interstellar gas, metallicity, andradiation. Chemical evolution equations, analogous to those usedto follow the chemical evolution of the Galaxy (e.g.[46,231,193]),are used to follow the evolution of the average stellar, gaseous,and radiative contents in each comoving volume in a self consis-tent manner. CCE models were pioneered by Pei and Fall [195],and most recently updated by Pei et al. [196]. Inputs parametersfor their model are the mean rest frame UV luminosity density asa function of redshift, and the mass of the ISM gas as determinedfrom H I column densities derived from studies of quasar absorp-tion lines through damped Lya systems. The decrease in the ISMgas with redshift and the UV luminosity density were used to de-rive a solution for the evolution of the CSFR with redshift whichis consistent with that determined from the extinction-correctedHa, and with SCUBA 850 and ISO 15 lm surveys. Similar to FE mod-els, population synthesis models were then used to calculate thestellar SED at each redshift, and an LMC extinction law wasadopted to calculate the fraction of starlight absorbed by the dust.A power-law distribution in dust temperature was used to calcu-late the spectrum of the reradiated IR emission. The model repro-duced various observational constraints, including the comovingrest-frame 0.44, 1.0, and 2.2 lm spectral luminosity densities inthe !0–2 redshift interval, the 12, 25, 60, and 100 lm local lumi-nosity densities; and the mean abundance of metals in dampedLya systems in the !0.4–3.5 redshift interval.

5.4. Semi-analytical (SA) models

SA models follow the formation and evolution of galaxies in acold dark matter Lambda dominated (KCDM) universe using thecosmological parameters derived from the 5-year Wilkinson Micro-wave Anisotropy Probe (WMAP5) observations [128] as the initialconditions. SA models then follow the growth and merger of darkmatter halos, and the emergence of galaxies which form as bary-onic matter falls into the potential wells of these halos. The fateof the infalling gas is determined by many different processes:the formation of stars in a multiphase interstellar medium, AGNand supernovae feedback processes that quench their formation,the evolution of the stellar radiation field, the heating and coolingof the interstellar medium and its chemical enrichment, the ex-change of material with the intergalactic medium through infalland galactic winds, and the growth of the central black hole. Adescription of recent developments and references to previouswork can be found in [222]. Model prediction are compared to abasic set of observational constraints such as the observed charac-teristics of galaxies: their morphology, colors, and spectral energydistribution, and morphology; and their integrated cosmologicalproperties: their number counts and luminosity function in differ-ent wavebands and redshifts, their mass function, the cosmic starformation rate, and the EBL generated by them. As in all EBL mod-els, determination of the galaxies’ SED is complicated by the de-tailed microscopic and large scale parameters needed to calculatethe amount of starlight that is absorbed by dust, and the spectrumof the reradiated emission. Recent SA models have combined themodels for galaxy formation with radiative transfer models todetermine the galaxies’ SED [103,102,222,241].

SA models are inherently complex, incorporating a large num-ber of physical processes, some poorly known, to derive galaxyproperties. However, they are the most physically motivated mod-els, and quite successful in reproducing a large number of observa-tional constraints.

5.5. Comparison of model predictions with observations

A detailed comparison of all model types with EBL limits andobservations was presented by Hauser and Dwek [124]. Here wewill represent mostly the models that have been developed sincethen: BE models by Stecker and Scully [229], Franceschini et al.[107], and Domínguez et al. [81]; The FE model of Finke et al.[99]; and the SA model of Gilmore et al. [117]. Fig. 9 comparesthe various models to the current limits and observations of theEBL. In general, all models, except for the BE models of Steckeret al. provide adequate fits to the EBL.

6. EBL constraints from c-ray observations of blazars

The attenuation of c-rays by the EBL can in principle be used todetermine the EBL intensity at wavelengths corresponding to thec-ray observations. Neglecting the possible scattering or produc-tion of second generation c-ray photons along the line of sight tothe blazar, the intrinsic c-ray flux from the blazar, FIntðEcÞ, can berelated to the observed one, FobsðEcÞ by:

FobsðEcÞ ¼ FIntðEcÞ exp½&sccðEcÞ' ð15Þ

where the optical depth, scc, is given by Eq. (9). Determination ofthe EBL assumes that all the attenuation is caused by interactionwith the EBL, instead of photons in or around the vicinity of the bla-zar. Furthermore, it requires knowledge of the intrinsic blazar spec-trum. Assuming that all the attenuation is attributed to the EBL,several upper limits have been derived on the EBL intensity by mak-ing various assumptions on the intrinsic blazar spectrum. The c-rayderived EBL limits are compared to those derived from UV to submillimeter observations in Fig. 10, and described below.

Fixed power law: Early observations of the blazars Mrk 421 and3C 279 suggested that their !GeV–TeV spectrum could be approx-imated by a single power law ([201,227], respectively). If so, thenany deviations of the observations from the extrapolated powerlaw to higher energies should be attributed to EBL attenuation.Stecker and de Jager [226] derived an upper limits at 1–5 lm of10 nW m&2 sr&1, assuming that a straight power law extrapolationof the spectrum of Mrk 421 from GeV energies obtained from theEGRET [163] with an index of C ¼ 1:96( 0:14 holds up to theTeV regime, where the index was measured by the Whipple Collab-oration to be C ¼ 2:25( 0:19. Biller et al. [61] included the statis-tical uncertainties of the GeV spectrum, and demonstrated vastlydifferent extrapolations with significantly higher upper limits,thereby yielding conservative upper limit to the EBL in themid-IR at 10 lm.

Fig. 9. Models of the EBL are compared to observational limits on the EBL.

124 E. Dwek, F. Krennrich / Astroparticle Physics 43 (2013) 112–133

m]µ [λ1 10 210

]-1

sr

-2) [

nW m

λ(ν Iν

1

10

2017 ICRC

Need to make some assumptions about EBL evolution with redshift…


Operating Gamma-ray Instruments

10

Fermi-LAT> Energy range: 100 MeV to > 300 GeV > Large duty-cycle > Large field of view

VERITAS

MAGIC

H.E.S.S.

HAWC

> Imaging Atmospheric Cherenkov Technique > E ~100 GeV to > 30 TeV > Precise energy &

angular reconstruction > Limited duty-cycle/FOV

> Water Cherenkov Technique > E ~1 - 100 TeV > Large duty-cycle > Large field of view

See S. Coutiño de Leon GA126 & GA255


Extragalactic Background Light with Fermi-LAT

11

formation of low-metallicity stars to be at z≥10 and its peak co-moving star-formation rate tobe lower than 0.5M⊙ Mpc−3 yr−1. This upper limit is already of the same order of the peakstar-formation rate of 0.2–0.6M⊙ Mpc−3 yr−1 proposed by (47) and suggests that the peakstar-formation rate might be much lower as proposed by (48).

Energy [GeV]210

γγτ

-110

1

10

LAT best fit -- 1 sigmaLAT best fit -- 2 sigmaFranceschini et al. 2008Finke et al. 2010 -- model CStecker et al. 2012 -- High OpacityStecker et al. 2012 -- Low OpacityKneiske et al. 2004 -- highUVKneiske et al. 2004 -- best fitKneiske & Dole 2010Dominguez et al. 2011Gilmore et al. 2012 -- fiducialAbdo et al. 2010

1.0≈z

Figure 1 Measurement, at the 68% and 95% confidence levels (including systematic uncertain-ties added in quadrature), of the opacity τγγ from the best fits to the Fermi data compared topredictions of EBL models. The plot shows the measurement at z≈1 which is the average red-shift of the most constraining redshift interval (i.e. 0.5≤z<1.6). The Fermi-LAT measurementwas derived combining the limits on the best-fit EBL models. The downward arrow representsthe 95% upper limit on the opacity at z=1.05 derived in (13). For clarity this figure shows onlya selection of the models we tested while the full list is reported in Table S1. The EBL modelsof (49), which are not defined for E≥250/(1 + z)GeV and thus could not be used, are reportedhere for completeness.

8

Energy [GeV]

γ γτ- e

210

0

0.5

1

z<0.2

210

0

0.5

1

0.2<z<0.5

10 210

0

0.5

1

0.5<z<1.6

Figure 2 Absorption feature present in the spectra of BL Lacertae objects as a function ofincreasing redshift (data points, from top to bottom). The dashed curves show the attenuationexpected for the sample of sources by averaging, in each redshift and energy bin, the opacitiesof the sample (the model of (7) was used) and multiplying this average by the best-fit scalingparameter b obtained independently in each redshift interval. The vertical line shows the criticalenergy Ecrit below which ≤5% of the source photons are absorbed by the EBL. The thin solidcurve represents the best-fit model assuming that all the sources have an intrinsic exponentialcut-off and that blazars follow the blazar sequence model of (32, 33).

9

> Source sample: 150 BL Lacs to z=1.6 (2FGL) > Fit observed spectra with

> b = 1.02 ± 0.23 (τγγ Franceschini 2008) > EBL detection @ ~6σ significance

Ackermann et al. 2012

✓dN

dE

◆

obs

=

✓dN

dE

◆

int

exp(�b⌧��

)

Assume only intrinsic cut-off


EBL with Recent Fermi-LAT Source Catalogues

12

2 Domınguez & Ajello

0 1 2 3 4 5 6 7 8

Spectral Index

0

5

10

15

20

25

30

35

40

Nu

mb

erof

Sou

rces

Γ2FHLint (This work)

Γ2FHLobs

(2FHL)

Figure 1. Distribution of observed (purple) and intrinsic (green)spectral index at energies greater than 50 GeV. The mean valuesof the distributions, calculated using the methodology followed byVenters & Pavlidou (2007), are shown with vertical lines.

2. DESCRIPTION OF THE DATA SAMPLE

The Fermi-LAT 2FHL catalog presents a total of 360sources over the whole sky detected at energies greaterthan 50 GeV. This catalog provides observed spectralindices, Γ2FHL

obs , which characterize the VHE power-lawspectra (since at these energies the LAT is not sensi-tive enough to detect spectral curvature) from 50 GeVto 2 TeV, redshifts, and synchrotron peak frequenciesfor 128 extragalactic sources. Using the synchrotronpeak frequency, we divide the sample into 33 low syn-chrotron peak blazars (LSP, log10(ν/Hz) ≤ 14) of which10 are FSRQs, 12 intermediate synchrotron peak blazars(ISP, 14 < log10(ν/Hz) < 15), and 83 high synchrotronpeak blazars (HSP, log10(ν/Hz) ≥ 15); therefore oursample contains mostly HSP rather than LSP and ISPblazars. To derive the intrinsic spectral indices for these128 sources, we modify the 2FHL analysis pipeline, whichis thoroughly described by Ackermann et al. (2015a), toinclude the EBL effect in the source models and fits. TheEBL effect is taken from the model by Domınguez et al.(2011a), which is derived from multiwavelength galaxyobservations. The data set and analysis parameters areexactly the same as in the 2FHL.3

Our analysis procedure leads to the histogram of spec-tral indices shown in Figure 1, where the observed andintrinsic index distributions are compared. In order toestimate the mean and spread of these distributions (af-fected by large uncertainties in the individual measure-ments), we follow the likelihood methodology describedby Venters & Pavlidou (2007). As expected, the maineffect of the EBL correction is to reduce the mean index(from 2.54 observed to 2.20 intrinsic). The mean intrin-sic index is relatively low, implying hard spectra. Theconfidence range calculated from the likelihood analysisis similar for the two distributions (σobs = 0.34 versusσint = 0.35).Ackermann et al. (2015a) show in their Figure 12 that

the observed indices evolve with redshift. Alternatively,in our Figure 2, we see no evidence for evolution withredshift of the intrinsic indices in our sample, which is

3 The intrinsic indices are provided for the Domınguez et al.(2011a) and Gilmore et al. (2012) Fiducial models in the 2FHLcatalog at http://fermi.gsfc.nasa.gov/ssc/data/access/lat/2FHL/

0.0 0.5 1.0 1.5 2.0

Redshift

−2

0

2

4

6

8

10

Γ2FHL

int

FSRQs

LSPs

ISPs

HSPs

Median values

Figure 2. The intrinsic spectral index versus redshift of our sam-ple. We note that only emission mechanisms and attenuation atthe sources contributes to this correlation, since the EBL effect isremoved. The sources are divided between FSRQs (red circles withblack edges), LSP blazars (red circles), ISP blazars (green squares),and HSP blazars (blue diamonds). Also, the median values (blackstars) and their 1σ uncertainties are shown. The uncertainty ofthe median is estimated from the distributions in each redshift binby bootstrapping, accounting for the uncertainties of the individ-ual measurements. The typical lower limit of 1.5 is shown with ahorizontal line.

dominated by HSP blazars. Thus, we see no significantevolution with redshift of the emission mechanisms inthis type of blazars. This conclusion needs to be inter-preted carefully because of potential bias in our sam-ple due to selection effects. In fact, at the higher red-shifts we are only sensitive to the most luminous blazarpopulations. We also see in Figure 2 that most of thesources (98%) are compatible within 1σ with having anindex larger than 1.5. This 1.5 is a fiducial value typi-cally used as a lower-limit index to derive upper limits onthe EBL spectral intensity based on local observations ofblazars (i.e., no EBL attenuation) and theoretical argu-ments of blazar photon emission (e.g., Aharonian et al.2006; Meyer et al. 2012). However, three sources haveextremely hard spectra (low index), more than 1σ lessthan 1.5. From Monte Carlo simulations, we find thatthese 3 sources are statistically expected from our dis-tribution of indices. These sources are listed in Table 1.Two of these sources, 1ES 0502+675 (Benbow 2011) andRBS 0413 (Aliu et al. 2012), have already been detectedby VERITAS.We also search in our sample for correlation between

the intrinsic index and the synchrotron peak frequencyas found in the 3LAC catalog (Ackermann et al. 2015b).However, no such correlation is found in our sample. Theprimary reason is that our sample is largely biased to-ward HSP blazars; therefore we are indeed missing ablazar population that is found at lower energies. Sec-ond, the spectral index scatter is larger at VHE thanat HE. This second effect is the result of poor photonstatistics together with the fact that at VHE we are sam-pling the decreasing side of the higher-energy peak of thebroadband spectral energy distribution.

3. SPECTRAL BREAKS

There are 122 sources with measured redshift in the2FHL catalog that are also found in the 3FGL catalog(Acero et al. 2015). For these sources, we complementour data with the observed HE spectral index provided

> 3LAC: 0.1 GeV - 100 GeV > 1st Fermi Hard Source Catalog: 10

GeV < E < 500 GeV > 2nd Fermi Hard Source Catalog: 50

GeV < E < 2 TeV > 3rd Fermi Hard Source Catalog: 10

GeV < E < 2 TeV > Softening visible for high energy

sources @ large z

Ajello et al. 2017 Domínguez & Ajello 2016

> Using 2nd Fermi Hard Source Catalog > Intrinsic indices using empirical models

(Domínguez 2011 & Gilmore 2012) > No spectral evolution with redshift

0.0 0.5 1.0 1.5 2.0 2.5Redshift

2

4

6

8

10

Ph

oton

Ind

ex

3FHL

2FHL

1FHL

3FHL/3LAC

0.0 0.5 1.0 1.5 2.0 2.5Redshift

−1

0

1

2

3

4

5

6

7

Γ3F

HL−Γ3F

GL

BL Lacs

FSRQs

BL Lacs+FSRQs

FSRQs

Fig. 16.— (Left panel) Observed spectral index vs. redshift of the 3LAC sources (energy range, 0.1–100GeV, gray circles), the median spectral index in some redshift bins of the 1FHL sources (10–500 GeV, bluecrosses), 2FHL sources (50–2000 GeV, orange stars), and 3FHL sources (10–2000 GeV, magenta circles).The uncertainties are calculated as the 68% containment around the median. The spectral index is seen todepend on the redshift at the energies where the EBL attenuation is significant. (Right panel) The differencebetween the 3FHL and 3FGL spectral index (∆Γ = Γ3FHL−Γ3FGL) over redshift for the BL Lac (blue crosses)and FSRQ (red circles) populations. The ∆Γ for both population types evolves similarly with redshift.

40

Single instrument


EBL with Blazar Spectral Energy Distribution Modeling

13

Domínguez et al. 2013> Model multiwavelength (radio to Fermi-LAT range) spectral energy distributions with synchrotron self-Compton model > Prediction for intrinsic/unabsorbed

spectra > Compare against spectra observed by

VERTAS/MAGIC/H.E.S.S.

⌧(E, z) = ln

✓✓dF

dE

◆

int

/

✓dF

dE

◆

obs

◆

> 15 blazars, low-state datasets with contemporaneous multiwavelength observations

> Calculate cosmic gamma ray horizon (τ=1) and compare to model of Domínguez 2011

Red: Domínguez 2011 Black: Best fit line


Multi-experiment measurement of the EBL

14

> Huge dataset: 30 blazars, 86 spectra >Measurements from H.E.S.S., MAGIC,

VERITAS, Whipple, HEGRA etc >Model-independent approach, with &

without lower limits from galaxy counts

11σ EBL detection, good consistency with several theoretical

models, rules out 4 models

Biteau & Williams 2015

m]µ [Wavelength1 10 210

]-1

sr

-2 [

nW m

EBL

inte

nsity

-110×81

2

3

4567

10

20

Model dependent constraints-LAT (2012)Fermi

H.E.S.S. (2013)1.13 x Gilmore+12

Model-independent constraints rays + localγ

rays onlyγ

Lower limits

Upper limits

m]µ [Wavelength1 10 210

]-1

sr

-2 [

nW m

EBL

inte

nsity

1

2

3

456

10

20

30

40Smooth modelSingle GaussianGaussian sum Redshift

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

[Te

V]En

ergy

Ran

ge

-110

1

10

BL LacsFSRQsUnderconstrained redshift

= 1τ

= 3τ

= 6τ


Redshift z1−10

Ener

gy [

TeV

]

1

10 = 1 τ

H.E.S.S. Preliminary

This work,

Systematic uncertainties

Franceschini et al. 2008

Dominguez et al. 2011

Finke et al. 2010

Gilmore et al. 2012 (Fiducial)

Kneiske & Dole 2010

Extragalactic Background Light with H.E.S.S.

15

m ]µWavelength [ 1 10 210

]-1

sr

-2In

tens

ity [

nW

m

1

10

]-3

Ener

gy d

ensi

ty [

eV

m

310

410

Energy [ eV ]2−101−101

H.E.S.S. Preliminary

This work,

Systematic uncertainties

H.E.S.S. 2013, model scaling

Meyer al. 2012 upper limit

rays onlyγVHE Biteau & Williams 2015,

Upper limits (Direct measurements)

Lower limits (Galaxy counts)

Franceschini et al. 2008

Dominguez et al. 2011

Finke et al. 2010

Gilmore et al. 2012 (Fiducial)

Kneiske & Dole 2010

> 9 blazars, 21 spectra (divide datasets into high and low states for variable sources)

> z=0.031-0.287 > Joint fit to spectral and EBL

parameters > EBL spectral energy distribution

divided into λEBL intervals, EBL intensity floating within intervals

> EBL detection @ 9.5σ significance

> Thorough treatment of systematic uncertainties > EBL evolution with

redshift > Energy scale

uncertainty > Choice of λEBL

intervals > Publication imminent

See A. Taylor GA150


Extragalactic Background Light with VERITAS

16

m]µ [λ1 10 210

]-1

sr

-2) [

nW m

λ(ν Iν

1

10

2017 ICRC

> 8 sources/spectra (no evidence for spectral variability for selected sources)

Target Source Class Redshift Exposure [hr] Observed Spectrum

1ES 2344+514 HBL 0.044 47 curvedH1426+428 HBL 0.129 81 curved

1ES 0229+200 HBL 0.14 140 power law1ES 1218+308 HBL 0.182 159 power law1ES 1011+496 HBL 0.212 37 power law

MS 1221.8+2452 HBL 0.218 2 power law1ES 0414+009 HBL 0.287 108 power lawPKS 1424+240 HBL 0.604 178 power law

> Generic EBL shapes (at z=0) from random spline generation > No constraints from galaxy counts or

direct measurements > Observed spectra corrected for EBL-

absorption (for appropriate energies and redshifts)


Extragalactic Background Light with VERITAS

17

m]µ [λ1 10

]-1 s

r-2

) [nW

mλ(

ν Iν

1

10

Upper limits/68% containment

Gilmore 2012 Fiducial

2017 ICRC

> EBL-corrected spectra fit with power law, curved models (if observed spectrum is curved) > Require concavity, spectral index > 1.5

> EBL shapes weighted by exp(-χ2/2) > Confidence intervals/upper limits extracted from probability distributions for

EBL intensity at fixed points in λEBL

> Preliminary, addition of 8 more blazars anticipated

m]µ [λ1 10 210

]-1 s

r-2

) [nW

mλ(

ν Iν

1

10

Upper limits/68% containment

Biteau&Williams 2015

2017 ICRC


Extragalactic Background Light with MAGIC

18

See A. Moralejo GA267 + wavelength-resolved approach

> 12 blazars, 32 spectra > z=0.03-0.944 > MAGIC-only constraints &

constraints using Fermi-LAT data

> τγγ Domínguez 2011 > Bin sources/EBL norm by redshift

✓dN

dE

◆

obs

=

✓dN

dE

◆

int

exp(�↵⌧��

)

MAGIC-only, relative to D11

¤  Compared to other gamma-based EBL scale measurements

¤  Good agreement with HESS and Fermi-LAT (EBL scale) measurements, and with VERITAS (note! : �- resolved approach)

19/07/17

18

A. Moralejo, Measurement of the EBL with MAGIC, 35th ICRC, Busan, Korea


Extragalactic Background Light with MAGIC

19

See A. Moralejo GA267 + wavelength-resolved approach

> Use Fermi-LAT spectral index & flux at pivot energy to constrain intrinsic spectrum > Contemporaneous Fermi-LAT

observations

> Mrk 421 and 1ES 1011+496 dominate combination for z<0.3

> EBL normalization consistent with 1MAGIC+Fermi, relative to D11

¤  With additional Fermi-LAT constraints, systematic uncertainties increase (+0.28 -0.26) as statistical ones decrease – under investigation

19/07/17

19

A. Moralejo, Measurement of the EBL with MAGIC, 35th ICRC, Busan, Korea


Single-source EBL Constraints: PKS 1441+25

20

6 The VERITAS Collaboration et al.

m ]µ[ Wavelength -110 -110×2 1 2 3 4 5 6

]-1

sr

-2[ n

W m

EBL

inte

nsity

3

4

5678

10

20 Biteau & Williams (2015) - 68% c.l.H.E.S.S. (2013) - 68% c.l.

-LAT (2012) - 68% c.l.Fermi

VERITAS - 95% c.l. upper limitPKS 1441-25 - cross section peakPKS 1441-25 - cross section FWHM

Lower limitsgalaxy counts

Upper limitsdirect observations

Fig. 3.— Near-ultraviolet to near-infrared spectrum of the EBL.The upper limit from this work is shown in blue, in regions corre-sponding to the peak and FWHM of the cross section (1 < ⌧ < 2).

5. DISCUSSION

The low energy threshold of VERITAS enabled the de-tection above 80GeV of one of the most distant VHEgamma-ray sources (z = 0.939), in a redshift range pre-viously accessible only to space-borne gamma-ray obser-vatories. We obtain stringent constraints on the EBLintensity below 1µm and conclude that galaxy surveyshave resolved most, if not all, of the sources of the EBL inthis region. This provides an excellent baseline for stud-ies above 1µm where the redshifted ultraviolet emissionof primordial stars could be detected (Dwek et al. 2005;Biteau & Williams 2015).The VHE detection of the highly-polarized source

PKS 1441+25 is contemporaneous with a period ofhard HE emission and of enhanced flux at all wave-lengths. The correlation between the radio, optical,and HE lightcurves, unusual for this class of sources(Max-Moerbeck et al. 2014b), together with slow multi-wavelength variability, suggest that the multi-band flarewas produced by a single region located ⇠ 104 � 105

Schwarzschild radii away from the black hole, which isconsistent with the VHE-gamma-ray escape condition.

PKS 1441+25 is by far the dimmest HE emitter of allVHE-detected FSRQs listed in the 3FGL catalog. WhileHE activity remains a prime trigger of VHE observa-tions, searches for new VHE-emitting quasars could alsofactor in radio-to-optical brightening and synchrotron-dominated X-ray emission, as reported for PKS 1441+25.These criteria will be of particular interest if appliedto distant FSRQs, possibly opening a new observationalwindow on the jets of blazars and on the transformationof the Universe’s light content with cosmic time.

This research is supported by grants from the U.S.Department of Energy O�ce of Science, the U.S. Na-tional Science Foundation and the Smithsonian Institu-tion, and by NSERC in Canada, with additional supportfrom NASA Swift GI grant NNX15AR38G. We acknowl-edge the excellent work of the technical support sta↵ atthe Fred Lawrence Whipple Observatory and at the col-laborating institutions in the construction and operationof the instrument. The VERITAS Collaboration is grate-ful to Trevor Weekes for his seminal contributions andleadership in the field of VHE gamma-ray astrophysics,which made this study possible.ASAS-SN thanks LCOGT, NSF, Mt. Cuba Astronom-

ical Foundation, OSU/CCAPP and MAS/Chile for theirsupport.The observations at Steward Observatory are funded

through NASA Fermi GI grant NNX12AO93G.CRTS is supported by the NSF grants AST-1313422

and AST-1413600.The OVRO 40-m monitoring program is sup-

ported in part by NASA grants NNX08AW31G andNNX11A043G, and NSF grants AST-0808050 and AST-1109911.The National Radio Astronomy Observatory is a fa-

cility of NSF operated under cooperative agreement byAssociated Universities, Inc.This research has made use of data from the MOJAVE

database that is maintained by the MOJAVE team (Lis-ter et al. 2009).

REFERENCES

Abdo, A. A., et al. 2010, Nature, 463, 919Acero, F., et al. 2015, ApJS, 218, 23Ackermann, M., et al. 2012, Science, 338, 1190Aharonian, F., et al. 2006, Nature, 440, 1018Aleksic, J., et al. 2011, ApJ, 730, L8Archambault, S., et al. 2014, ApJ, 785, L16Atwood, W. B., et al. 2009, ApJ, 697, 1071Barnacka, A., et al. 2014, A&A, 567, A113Biteau, J., & Williams, D. A. 2015, accepted in ApJ,

arXiv:1502.04166Bottcher, M. 2008, in American Institute of Physics Conference

Series, ed. F. A. Aharonian, W. Hofmann, & F. Rieger, Vol.1085, 427–430

Cerruti, M., et al. 2013, ApJ, 771, L4Domınguez, A., et al. 2011, MNRAS, 410, 2556Donea, A.-C., & Protheroe, R. J. 2003, Astroparticle Physics, 18,

377Drake, A. J., et al. 2009, ApJ, 696, 870—. 2013, ApJ, 763, 32Dwek, E., Arendt, R. G., & Krennrich, F. 2005, ApJ, 635, 784Fossati, G., et al. 1998, MNRAS, 299, 433Franceschini, A., Rodighiero, G., & Vaccari, M. 2008, A&A, 487,

837Gehrels, N., et al. 2004, ApJ, 611, 1005

Gilmore, R. C., et al. 2012, MNRAS, 422, 3189Harrison, F. A., et al. 2013, ApJ, 770, 103H.E.S.S. Collaboration et al. 2013a, A&A, 554, A107H.E.S.S. Collaboration, et al. 2013b, A&A, 550, A4Holder, J. 2011, International Cosmic Ray Conference, 12, 137Jenkins, E. B., & Savage, B. D. 1974, ApJ, 187, 243Kalberla, P. M. W., et al. 2005, A&A, 440, 775Kaspi, S., et al. 2007, ApJ, 659, 997Lister, M. L., et al. 2009, AJ, 138, 1874—. 2011, ApJ, 742, 27Madau, P., & Pozzetti, L. 2000, MNRAS, 312, L9MAGIC Collaboration et al. 2008, Science, 320, 1752Max-Moerbeck, W., Richards, J. L., Hovatta, T., Pavlidou, V.,

Pearson, T. J., & Readhead, A. C. S. 2014a, MNRAS, 445, 437Max-Moerbeck, W., et al. 2014b, MNRAS, 445, 428Mirzoyan, R. 2015, The Astronomer’s Telegram, 7416, 1Moderski, R., et al. 2005, MNRAS, 363, 954Pacciani, L. 2015, The Astronomer’s Telegram, 7402, 1Richards, J. L., et al. 2011, ApJS, 194, 29Roming, P. W. A., et al. 2005, Space Sci. Rev., 120, 95Sasada, M., et al. 2014, ApJ, 784, 141Schmidt, G. D., Stockman, H. S., & Smith, P. S. 1992, ApJ, 398,

L57Shappee, B. J., et al. 2014, ApJ, 788, 48


[ MeV ]Energy -1010 -810 -610 -410 -210 1 210 410 610

]-1 s-2

[ e

rg c

mν

Fν

Ener

gy fl

ux

-1410

-1310

-1210

-1110

-1010 April 2015VERITAS

-LATFermiNuSTARSwiftASAS-SN

Other observations - May 2015Swift - January 2015Swift - June 2010Swift

Archival (ASDC)

Broad-band modelingSynchrotronSelf ComptonExternal Compton(torus)

[ Hz ]Frequency 1110 1310 1510 1710 1910 2110 2310 2510 2710

]-1 [

erg

sν

Lν

Lum

inos

ity

4410

4510

4610

4710

[ MeV ]Energy 410×4 510 510×2 510×3

]-1

s-2

[ e

rg c

mν

Fν

Ener

gy fl

ux

-1210

-1110

-1010

VERITAS -LATFermi

Corrected for EBL

Observed

[ MeV ]Energy -310 -210

]-1

s-2

[ e

rg c

mν

Fν

Ener

gy fl

ux

-1310×4

-1310×5

-1310×6

-1310×7

-1310×8

-1210

-1210×2NuSTAR -XRTSwift

Fig. 2.— Multiwavelength emission of PKS 1441+25. Side panels show the X-ray (top) and gamma-ray emission (bottom) in April 2015(MJD 57133-57140). The various exposures and the model are discussed in Sec. 2 and 3, respectively.

monitored PKS 1441+25 at 15GHz since late 2009. A15GHz VLBA image obtained by the MOJAVE program(Lister et al. 2009) on 2014 March 30 (MJD 56381) showsa compact core and a bright, linearly polarized jet fea-ture located 1.2milliarcsec downstream, at position angle�68 �. Both features have relatively high fractional po-larization (⇠ 10%), and electric vectors aligned with thejet direction, at an angle of 102 � similar to that measuredby SPOL, indicating a well-ordered transverse magneticfield. The fractional polarization level of the core fea-ture is among the highest seen in the MOJAVE program(Lister et al. 2011).The 2008-2015 observations of PKS 1441+25 shown in

Fig. 1 reveal a brightening of the source in the radio,optical, and HE bands starting around MJD 56900. Asimple Pearson test (see caveats in Max-Moerbeck et al.2014a) applied to the radio and HE long-term lightcurvesshows a correlation coe�cient r = 0.75 ± 0.02, di↵eringfrom zero by 5.4� based on the r-distribution of shu✏edlightcurve points. Similarly, the analysis of the opticaland HE lightcurves yields r = 0.89± 0.02, di↵ering fromzero by 4.8�. The discrete correlation functions displaybroad, zero-centered peaks with widths of ⇠ 100 days, in-dicating no significant time lags beyond this time scale.During the period marked by gray dashed lines in Fig. 1,observations on daily timescales from optical wavelengthsto X-rays reveal fractional flux variations smaller than25%, compatible with the upper limits set by Fermi-LATand VERITAS (30% and 110% at the 95% confidencelevel, respectively). Such flux variations are small withrespect to the four orders of magnitude spanned in ⌫F⌫ ,enabling the construction of a quasi-contemporaneousspectral energy distribution in Sec. 3.

3. EMISSION SCENARIO

The spectral energy distribution, with the X-ray-to-VHE data averaged over the active phase in April 2015(MJD 57133-57140), is shown in Fig 2. The optical-to-X-ray spectrum is well described by a power law with

photon index � = 2.29 ± 0.01 from 2 eV to 30 keV, in-cluding a 10% intrinsic scatter in the fit procedure thataccounts for the small-amplitude optical-to-UV variabil-ity. This spectrum suggests a single synchrotron compo-nent peaking below 2 eV ⇠ 5 ⇥ 1014 Hz, created by anelectron population of index p = 2� � 1 ⇠ 3.58 ± 0.02.As expected in FSRQs (Fossati et al. 1998), the emissionof PKS 1441+25 is dominated by the gamma-ray com-ponent, well-described by a single component peaking at3.3+1.8

�1.1 GeV.The detection of gamma rays up to 200GeV, about

400GeV in the galaxy’s frame, suggests that the emit-ting region is located beyond the BLR, or else pair pro-duction would suppress any VHE flux even for a flatBLR geometry (Tavecchio & Ghisellini 2012). The el-evated radio state, correlated with the optical and HEbrightening, also suggests synchrotron emission outsideof the BLR where synchrotron self-absorption is smaller.The hypothesis of large-scale emission is strengthened bythe week-long duration of the optical-to-gamma-ray flare.This behavior contrasts with other observations of brightFSRQs, displaying di↵erent flux variations at di↵erentwavelengths (e.g. Abdo et al. 2010), more in line withmulti-component scenarios. The flare of PKS 1441+25appears to be one of the few events whose detailed tem-poral and spectral multiwavelength features are consis-tent with the emission of a single component beyond theBLR.The BLR size can be derived using the estimated black-

hole mass, MBH = 107.83±0.13M� (Shaw et al. 2012),assuming rBLR ' 1017 cm ⇥

pLdisk/1045 erg s�1 (Kaspi

et al. 2007) and an accretion disk luminosity that is afraction ⌘ = 10% of the Eddington luminosity. Alterna-tively, Ldisk can be estimated from the BLR luminosity asLdisk ' 10 ⇥ LBLR, with LBLR = 1044.3 erg s�1 (Xiong& Zhang 2014). Both estimates yield rBLR ' 0.03 pc,setting a lower limit on the distance between the blackhole and the emitting region of r & 5, 000 Schwarzschild

> Detected by MAGIC and VERITAS in April 2015 > z=0.939


by MAGIC (Mirzoyan 2015).60

2. OBSERVATIONS OF PKS 1441+2561

PKS 1441+25 was detected from 2015 April 2162

(MJD 57133) to April 28 (MJD 57140) with VERI-63

TAS, an array of four imaging atmospheric Cherenkov64

telescopes located in southern Arizona (Holder 2011).65

VERITAS imaged gamma-ray showers from the source66

above 80GeV, enabling the detection of PKS 1441+2567

(VER J1443+250) at a position consistent with its ra-68

dio location, with a significance of 7.7 standard devi-69

ations (�) during the 15.0 h exposure (2710 ON-source70

events, 13780 OFF-source events, OFF normalization of71

1/6). Using a standard analysis procedure with cuts op-72

timized for low-energy showers (Holder et al. 2006), we73

measure an average flux �(> 80GeV) = (5.0 ± 0.7) ⇥74

10�11 cm�2 s�1 with a photon index �VHE = 5.3 ± 0.575

up to 200GeV. The day-by-day lightcurve is com-76

patible with constant emission in that period. Sub-77

sequent observations in May (MJD 57155-57166, 3.8 h78

exposure) showed no significant excess (660 ON-source79

events, 3770 OFF-source events, OFF normalization of80

1/6), resulting in a flux upper limit �(> 80GeV) <81

4.3⇥10�11 cm�2 s�1 at the 99% confidence level. These82

results have been cross checked with an independent cal-83

ibration and analysis chain (Cogan 2008). Monte-Carlo84

simulations indicate systematic uncertainties on the VHE85

energy scale and photon index of 20% and 0.2, respec-86

tively, for this set of observations. The systematic uncer-87

tainty on the flux of this source is estimated to be 60%,88

including the uncertainty on the energy scale.89

The LAT pair-conversion telescope onboard the Fermi90

satellite has surveyed the whole sky in the high-energy91

(HE, 100MeV < E < 100GeV) band since August 200892

(Atwood et al. 2009). We analyzed the LAT data using93

the public science tools v10r0p5 (Pass-8) with a region of94

interest and source region of 10 � and 20 � radius, respec-95

tively. We reconstruct the spectrum of PKS 1441+25 be-96

tween 100MeV and 100GeV in four-week (MJD 54705-97

57169) and two-week (MJD 57001-57169) bins assuming98

a power-law model with a free normalization and pho-99

ton index (purple points in Fig. 1), as well as in one-100

day bins (pink points in Fig. 1) fixing the photon in-101

dex to its best-fit average value in MJD 57001-57169,102

�HE = 1.97 ± 0.02, slightly harder than reported in the103

3FGL, 2.13±0.07, for the 2008-2012 period (Acero et al.104

2015). The source is in a high state during MJD 57001-105

57169 (see Fig. 1), with 100MeV�100GeV integral-flux106

values one to two orders of magnitude above the 3FGL107

value, (1.3± 0.1)⇥ 10�8 cm�2 s�1. During the two-week108

period contemporaneous with the VERITAS detection,109

the source shows a flux of 34±3 (same units) and a hard110

index of 1.74 ± 0.05. Although a power-law model is111

used for robustness in the lightcurve determination, the112

spectrum shows a hint of curvature during MJD 57127-113

57141, with a log parabola preferred at the 3.4� level114

with respect to a power law.115

X-ray observations with the NuSTAR and Swift satel-116

lites were triggered following the VHE detection. NuS-117

TAR, a hard-X-ray instrument sensitive to 3 � 79 keV118

photons (Harrison et al. 2013), observed the source on119

MJD 57137 for an exposure time of 38.2 ks. The data120

were reduced using the NuSTARDAS software v1.3.1. Swift-121

120 130 140 150 160

deg

ang

le

Pola

rizat

ion

95100105110115120

SPOL

120 130 140 150 160

% fr

actio

n Po

lariz

atio

n

2025303540

SPOL

120 130 140 150 160

-1Å

-1 s

-2er

g cm

-15

x10

(O-U

V)Φ 0

0.51

1.52

2.5 VBU

W1M2W2

V - ASAS-SNV - Catalina

-UVOT / ASAS-SN / CatalinaSwift

120 130 140 150 160

-1 s

-2 e

rg c

m-1

210

(2-1

0 ke

V)Φ

00.5

11.5

2

-XRTSwift

120 130 140 150 160

0.1-

100

GeV

I

ndex

1.71.81.9

22.12.2

-LATFermi

120 130 140 150 160

-1 s

-2 c

m-8

10

(0.1

-100

GeV

)Φ 0

20406080

-LATFermi

120 130 140 150 160

-1 s

-2 c

m-1

110

(> 8

0 G

eV)

Φ -505

1015

VERITAS

20 40 60 80 100 120 140 160

0.1-

100

GeV

I

ndex

1.71.81.9

22.12.22.3

-LATFermi

20 40 60 80 100 120 140 160

-1 s

-2 c

m-8

10

(0.1

-100

GeV

)Φ 0

50100150200

-LATFermi

-2000 -1500 -1000 -500 0

-1 s

-2 c

m-8

10

(0.1

-100

GeV

)Φ 0

102030405060

-LATFermi-2000 -1500 -1000 -500 0

-1Å

-1 s

-2er

g cm

-15

x10

(V)

Φ 00.5

11.5

2 Catalina-2000 -1500 -1000 -500 0

Jy

(15

GHz

)Φ

00.10.20.30.40.50.60.7

OVRO

MJD - 57000-2000 -1500 -1000 -500 0

MJD - 5700020 40 60 80 100 120 140 160

MJD - 57000120 130 140 150 160

Fig. 1.— Top: observations from 2008 to 2015. Middle: obser-vations from December 2014 to May 2015. Bottom: observationsin April and May. The gray dashed lines mark the VHE discoveryby MAGIC and the end of VERITAS observations in April.

Abeysekara et al. 2015

> Power-law extrapolation from HE range, τγγ Gilmore 2012

> Competitive with multi-source constraints in near-UV to near-IR

✓dN

dE

◆

obs

=

✓dN

dE

◆

int

exp(�↵⌧��

)


Projections for the Cherenkov Telescope Array

21

> 10x improvement in sensitivity over existing instruments > 118 telescopes, 3 telescope

sizes, northern and southern site

> E ~20 GeV to 300 TeV

> Improvement in sensitivity → better spectral measurements on known sources > Broad energy coverage → simultaneous measurement of intrinsic and

absorbed spectra > Extragalactic survey

> 25% of sky > Increase population of blazars usable for EBL studies

> Transient program > Rapid follow-up on flaring gamma-ray sources > Detect flaring blazars to z~1 > Gamma-ray bursts

See R. Ong GA325 and T. Hassan GA145

(TeV)R

Energy E2−10 1−10 1 10 210

)-1 s

-2 x

Flu

x Se

nsiti

vity

(erg

cm

2 E 13−10

12−10

11−10

ww

w.c

ta-o

bser

vato

ry.o

rg (2

017-

07-0

5)

CTA South

50 h

Differential flux sensitivity

LAT Pas

s 8 (1

0y, (l

,b)=(0

,0))

LAT Pas

s 8 (1

0y, (l

,b)=(1

20,45

))

MAGIC 50 h VERITAS 50 h

H.E.S.S. 50 h

HAWC 1 yearHAWC 5 year

CTA North 5

0 h

Differential flux sensitivity


Projections for the Cherenkov Telescope Array

22

Redshift z0 0.1 0.2 0.3 0.4 0.5

αN

orm

aliz

ed E

BL

opac

ity

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

H.E.S.S. 2013

-LAT 2012Fermi

CTA 2020-30

PRELIMINARY

See F. Gaté GA185 + projections for studies of intergalactic magnetic field, Lorentz invariance violation,

axion-like particles

> Projections based on known sources in low flux state > 15 sources, 100 hours exposure on each

Projected ~10% uncertainty on EBL normalization to z=0.5

> Intrinsic spectra extrapolated from Fermi 3FHL catalog (power laws & log parabolas, spectral cutoffs @ 1 TeV), τγγ Domínguez 2011

> Bin sources/EBL normalization by redshift

✓dN

dE

◆

obs

=

✓dN

dE

◆

int

exp(�↵⌧��

)


Conclusions.

> Properties of extragalactic light interesting in their own right, important for interpretation of blazar spectral observations

> Indirect measurement of extragalactic background light with gamma-ray observations continues to be an active area > Many new results! > Operating experiments continue to improve measurements/constraints > Cherenkov Telescope Array promises exciting improvements

> Not mentioned in this talk: EBL is important for other studies related to gamma-ray propagation from distant emitters > Intergalactic magnetic field > Axion-like particle searches > Lorentz invariance violation (through modification of threshold for pair

production) > Measurement of Hubble constant

> Extremely rich topic!

23

Photo: M. Fernandez


Backup.

24


Source Selection: MAGIC

25

PoS(ICRC2017)604

Measurement of the EBL with MAGIC A. Moralejo Olaizola

1. Introduction

The Extragalactic Background Light (EBL) is a cosmic diffuse radiation field that enclosesessential information about galaxy evolution and cosmology. It is mainly composed by ultravio-let, optical, and near-infrared light emitted by stars and its re-emission to longer wavelengths byinterstellar dust, which produces its characteristic double peak spectral energy distribution. This ra-diation is accumulated over the cosmic history and redshifted by the expansion of the Universe (see[1] for a review, and references therein). Other contributions to the EBL may exist such as thosecoming from the accretion on super-massive black holes, light from the first stars, or even moreexotic sources such as products of the decay of relic dark matter particles. The actual contributionof these components to the total background is poorly known. The direct photometric detection ofthe EBL is challenging because of strong foreground, mainly zodiacal light but also the brightnessof our own Galaxy. Therefore, attempts of direct detection are subject to large uncertainties andbiases. Other detection analyses focus on measuring the background anisotropies, which still pro-vides inconclusive results (e.g. [2, 3]). These techniques cannot provide any information about thefundamental EBL evolution.

An alternative methodology to estimate the EBL is based on counting photons in different pho-tometric bands using deep-galaxy-surveys data (e.g. [4]). This procedure results in EBL estimatesthat can be considered robust lower limits. However, cosmic variance may contribute to systematicuncertainties using this technique. Also light from the outer regions of normal galaxies or stars infaint undetected galaxies can be missed.

There are other efforts centered on building empirical models using different complementarymethodologies. Following the classification by [5] (see references therein), these models are di-vided in four different classes: (1) Forward evolution models that use semi-analytical models ofgalaxy formation, (2) Backward evolution models based on local or low redshift galaxy data, whichare extrapolated to higher redshifts making some assumptions on the galaxy evolution, (3) Inferred

Source type redshift period observationtime (h)

Markarian 421 (15 spectra) HBL 0.030 20130410 - 19, 20140426 43.81ES 1959+650 HBL 0.048 20151106 - 18 4.8OT 546 (1ES 1727+502) HBL 0.055 20151012 - 20151102 6.4BL Lacertae IBL 0.069 20150615 1.01ES 0229+200 HBL 0.14 2012 - 2015 105.21ES 1011+496 HBL 0.212 20140206 - 20140307 11.8PKS 1510-089 (2 spectra) FSRQ 0.361 20150518-19, 20160531 5.0PKS 1222+216 FSRQ 0.432 20100618 0.5PG 1553+113 (5 spectra) HBL 0.43 - 0.58 2012 - 2016 66.4PKS 1424+240 (2 spectra) HBL 0.604 2014 - 2015 49.1PKS 1441+25 FSRQ 0.939 20150418 - 20150423 20.1QSO B0218+35 FSRQ 0.944 20140725 - 20140726 2.1

Table 1: List of the 32 MAGIC spectra used in the determination of the EBL density.

2


Source Selection: CTA

26

1ES 0229+200 H 1426+428 1ES 1101-232 Mrk 421 Mrk 501 PKS 2155-304 1ES 1011+496 3C 66A W Comae AP Librae BL Lac PKS 1510-089 PKS 1222+216 M 87 NGC 1275

z = 0.140 z = 0.129 z = 0.186 z = 0.031 z = 0.034 z = 0.116 z = 0.212 z = ~0.4 z = 0.102 z = 0.049 z = 0.069 z = 0.361 z = 0.435 z = 0.0044 z = 0.018

Documents

The Extragalactic Background Light · The Extragalactic Background Light. Elisa Pueschel ICRC 2017, Busan, South Korea 19.07.2017 Constraints from TeV Blazar Observations H. Dole