The Detection of the Cosmic γ-ray Horizon The Detection of thephysics.ucsc.edu/~joel/Primack-EBL-Barcelona.pdf · Fourth International Fermi Symposium, Monterey, CA October 28 –

EBLJoel Primack & Alberto Domínguez

Exploring the Non-thermal Universe with Gamma RaysOn the occasion of the 60th birthday of Felix AharonianBarcelona, November 6 - 9, 2012

Exploring the Non-thermal Universe with Gamma RaysOn the occasion of the 60th birthday of Felix Aharonian

Barcelona, November 6 - 9, 2012

ExtragalacticBackground Light

The Detection of theThe Detection of the

Cosmic Cosmic γγ-ray Horizon-ray Horizon

Alberto Domínguez(University of California, Riverside)

Collaborators:

Justin Finke, Francisco Prada, Joel Primack, Brian Siana

Fourth International Fermi Symposium, Monterey, CA

October 28 – November 2, 2012




Collaborators:







Collaborators:




Data from (non-) attenuation of gamma rays from blazars and gamma ray bursts (GRBs) give upper limits on the EBL from the UV to the mid-IR that are only a little above the lower limits from observed galaxies. New data on attenuation of gamma rays from blazers now lead to statistically significant measurements of the cosmic gamma ray horizon (CGRH) as a function of source redshift and gamma ray energy that are independent of EBL models. These new measurements are consistent with recent EBL calculations based both on multiwavelength observations of thousands of galaxies and also on semi-analytic models of the evolving galaxy population. Such comparisons account for all the light, including that from galaxies too faint to see. Catching a few high-redshift GRBs with Fermi or low-threshold atmospheric Cherenkov telescope (ACT) arrays could provide important new constraints on the high-redshift star formation history of the universe.

Joel Primack & Alberto Domínguez

Extragalactic Background Light (EBL)

Gamma Ray Attenuation due to γγ → e+e-

If we know the intrinsic spectrum, we can infer the optical depth τ(E,z) from the observed spectrum. In practice, we typically assume that dN/dE|int is not harder than E-Γ with Γ = 1.5, since local sources have Γ ≥ 2. More conservatively, we can assume that Γ ≥ 2/3.

Illustration: Mazin & Raue

Local EBL data and Local EBL data and γ-ray γ-ray limitslimits

Local EBL ObservationsΓ ≥ 1.5Γ ≥ 2/3

Evolution Calculated from ObservationsUsing AEGIS Multiwavelength Data

Alberto DomÍnguez, Joel Primack, et al. (MNRAS, 2011)

http://aegis.ucolick.org/

0.7 ☐°

http://aegis.ucolick.org

http://aegis.ucolick.org

χ SED FittingLe PHARE code for fitting the SWIRE templates in FUV, NUV, B, R, I, Ks, IRAC1, 2, 3, 4 and MIPS24

Domínguez+ 11

2

Quiescent

Star-forming Starburst AGN-type

Best SE

D Fits

Worst SE

D Fits

SED-Type EvolutionLocal fractions, z<0.2:

Goto+ 03, morphologically classified from Sloanconverted to spectral classification using resultsfrom Galaxy Zoo Skibba+ 09 ~6% blue ellipticals Schawinski+ 09 ~25% red spirals

Results: 35% red-type galaxies65% blue-type galaxies

High-redshift universe, z>1:

Two approaches:1. Keep constant the fractions of our last redshift bin (Fiducial Model), or2. Quickly increase starburst population from 16% at z = 0.9 to 60% at z ≥ 2

We find that the differences in the predicted EBL are small except at long wavelengths, affecting attenuation only for E ≥ 5 TeV. Domínguez+11

Maximum uncertainty due to photometry and fit errors

Local Luminosity Density

Domínguez+11

Local EBL: data, Local EBL: data, γ-ray γ-ray limits, and modelslimits, and models

Domínguez+ 11Domínguez+ 11

Γ ≥ 1.5Γ ≥ 2/3

Local EBL Observationsvs. Domínguez+11

Propagating errors in SED fitsand redshift extrapolation

When we first tried doing this (Primack & MacMinn 1996, presented at Felix Aharonian’s first Heidelberg conference), both the stellar initial mass function (IMF) and the values of the cosmological parameters were quite uncertain. After 1998, the cosmological model was known to be ΛCDM although it was still necessary to consider various cosmological parameters in models. Now the parameters are known rather precisely, and our latest semi-analytic model (SAM) uses the current (WMAP5) cosmological parameters. With improved simulations and better galaxy data, we can now normalize SAMs better and determine the key astrophysical processes to include in them.

Remaining uncertainties include whether the IMF is different in different galaxies (possibly “bottom-heavy” in massive galaxies), feedback from AGN, the nature of sub-mm galaxies, and the star formation rate at high redshifts.

EBL Calculated by Forward Evolution using SAMs

~1012

z=5.7 (t=1.0 Gyr)

z=1.4 (t=4.7 Gyr)

z=0 (t=13.6 Gyr)

Springel et al. 2005 Wechsler et al. 2002

• cosmological parameters are now well constrained by observations

• mass accretion history of dark matter halos isrepresented by ‘merger trees’ like the one at left

Present status of ΛCDM“Double Dark” theory:

time

Forward Evolution

Cluster Data

Galaxy Formation in ΛCDM• gas is collisionally heated when perturbations ‘turn

around’ and collapse to form gravitationally bound structures

• gas in halos cools via atomic line transitions (depends on density, temperature, and metallicity)

• cooled gas collapses to form a rotationally supported disk

• cold gas forms stars, with efficiency a function of gas density (e.g. Schmidt-Kennicutt Law)

• massive stars and SNae reheat (and in small halos expel) cold gas and some metals

• galaxy mergers trigger bursts of star formation; ‘major’ mergers transform disks into spheroids and fuel AGN

• AGN feedback cuts off star formation

White & Frenk 91; Kauffmann+93; Cole+94;Somerville & Primack 99; Cole+00; Somerville, Primack, & Faber 01; Croton et al. 2006; Somerville +08; Fanidakis+09; Guo+2011; Somerville, Gilmore, Primack, & Domínguez 12 (discussed here)

z=0 Luminosity DensityEvolving Luminosity Density

Some Results from our Semi-Analytic Models

WMAP1WMAP5

Gilmore, Somerville, Primack, & Domínguez (2012)

Modelling of the EBL and gamma-ray spectra 3193

Figure 2. Left: the luminosity density of the local universe. The solid black line is the WMAP5 model, and the dotted line is the C!CDM model. Data at anumber of wavelengths are shown from GALEX (blue circles), SDSS (red stars; Montero-Dorta & Prada 2009), 6dF (light blue squares; Jones et al. 2006),2MASS (green stars; Cole et al. 2001; Bell et al. 2003). In the mid- and far-IR, the orange squares are from IRAS (Soifer & Neugebauer 1991), while bluestars are from an analysis of local emissivity using data from IRAS, ISO and SCUBA (Takeuchi et al. 2001). Right: three-dimensional representation of theevolution of the luminosity density in our WMAP5 model as a function of wavelength and redshift.

and integrated counts of galaxies. Direct measurements provide anabsolute measurement of the background light without regard to thesources responsible, but require subtraction of foreground sourcespresent in the Milky Way and our Solar system in order to isolatethe extragalactic signal. Integration of galaxy counts (galaxies perunit sky area at a given magnitude) is a way to set firm lower limitson the EBL, although the degree to which these measurementsconverge on the true value often remains controversial. The fluxfrom faint sources will converge mathematically if the slope ofthe counts plotted on a log number versus flux diagram is flatterthan unity, or in terms of magnitudes if " < 0.4, for ln (N) ! " m.As expounded by Bernstein (2007), photometry of faint galaxiesis fraught with difficulty in untangling the faint galactic fringesfrom the background, and it is possible to miss 50 per cent or moreof the light associated with extended sources in simple aperturephotometry.

Large-scale surveys such as the SDSS, the 6-degree Field survey(6dF) and the 2MASS have provided us with an accurate accountingof the galaxies in the local universe, and surveys with the HSThave complemented this data with extremely deep counts. Satelliteinstruments such ISOCAM, IRAC and MIPS provide data in themid- and far-IR. A detailed presentation of galaxy number countsin our models compared with data can be found in SGPD12.

Our prediction for the local EBL is generally in agreement withlower limits from integrated number counts. In the UV, limits fromGardner, Brown & Ferguson (2000) are considerably higher thanthe measurement from GALEX (Xu et al. 2005). This may be ex-plained by the former’s use of data from the balloon-based FOCAexperiment to find bright counts, which were in disagreement withthose from GALEX at several magnitudes. Preliminary Herschelcounts data from Berta et al. (2010) set only a weak lower limit onthe far-IR background peak, and the author acknowledges that onlyabout half the total IR background is likely being resolved.

Absolute measurements of the EBL require the removal of fore-ground sources, including stars, ISM emission and sunlight reflected

from dust in the inner Solar system (often called ‘zodiacal’ light).The most robust direct measurements of the IR background to datecome from DIRBE and FIRAS instruments on the Cosmic Back-ground Explorer (COBE) satellite, though they are still fraught withuncertainty in sky subtraction (see fig. 2 in Hauser & Dwek 2001).The near-IR flux has been calculated from DIRBE observationsby a variety of authors (Gorjian, Wright & Chary 2000; Wright &Reese 2000; Cambresy et al. 2001; Wright 2001; Levenson, Wright& Johnson 2007) using foreground source subtraction techniquesand modelling of the zodiacal light, and has generally yielded highestimates in this range compared to number counts. Another notableattempt to measure the near-IR background was Levenson & Wright(2008), which used IRAC data to calculate the best-fitting flux at3.6 µm using a profile-fit to estimate the light from the unobservablefaint fringes of galaxies. These results were 70 per cent higher thanthose of the aperture method of Fazio et al. (2004), highlightingthe large uncertainties that galaxy fringe issues can bring to EBLmeasurement.

The present-day EBL obtained in each of our models is shownin Fig. 4. We also show results from D11 for comparison. Thelocal EBL is calculated by integrating over the luminosity densityat all wavelengths beginning at z = 7.5, and accounting for theredshifting and dilution of photons as the universe expands. TheEBL at a redshift z0 and frequency #0 in proper coordinates can bewritten as (Peebles 1993)

J (#0, z0) = 14!

! "

z0

dl

dz

(1 + z0)3

(1 + z)3$(#, z) dz, (1)

where $(#, z) is the galaxy emissivity at redshift z and frequency# = #0(1 + z)/(1 + z0), and dl/dz is the cosmological line element,which isdl

dz= c

(1 + z)H0

1"%m(1 + z)3 + %!

(2)

for a flat !CDM universe. We assume here that the EBL pho-tons evolve passively after leaving their source galaxies and are

C# 2012 The Authors, MNRAS 422, 3189–3207Monthly Notices of the Royal Astronomical Society C# 2012 RAS

18 20 22 24-6

-5

-4

-3

-2

-1

18 20 22 24 26-6

-5

-4

-3

-2

-1

18 20 22 24 26-6

-5

-4

-3

-2

-1

18 20 22 24 26-6

-5

-4

-3

-2

-1

Evolving Luminosity Functions

18 20 22 24-6

-5

-4

-3

-2

-1

18 20 22 24-6

-5

-4

-3

-2

-1

18 20 22 24-6

-5

-4

-3

-2

-1

18 20 22 24-6

-5

-4

-3

-2

-1

B-band


K-band


An advantage of the SAM approach is that it is possible to compare predictions and observations at all redshifts and in all spectral bands.

3.6, 8, 24 and 24, 70, 160, & 850 μm Bands


Somerville, Gilmore, Primack, & Domínguez (2012)

Number Counts in UV, b, v, i, and z Bands

Worst failure is at 850 μm

EBL from our Semi-Analytic Models



Figure 4. The predicted z = 0 EBL spectrum from our fiducial WMAP5 model (solid black) and WMAP5+fixed (dash–dotted violet) dust parameters, andC!CDM (dotted black) models, compared with experimental constraints at a number of wavelengths. D11 is shown for comparison in dashed–dotted red withthe shaded area indicating the uncertainty region. Data: upward pointing arrows indicate lower bounds from number counts; other symbols are results fromdirect detection experiments. Note that some points have been shifted slightly in wavelength for clarity. Lower limits: the blue–violet triangles are results fromHST and Space Telescope Imaging Spectrograph (STIS; Gardner et al. 2000), while the purple open triangles are from GALEX (Xu et al. 2005). The solid greenand red triangles are from the Hubble Deep Field (Madau & Pozzetti 2000) and Ultra Deep Field (Dolch & Ferguson, in preparation), respectively, combinedwith ground-based data, and the solid purple triangle is from a measurement by the Large Binocular Camera (Grazian et al. 2009). In the near-IR J, H and Kbands, open violet points are the limits from Keenan et al. (2010). Open red triangles are from IRAC on Spitzer (Fazio et al. 2004), and the purple triangle at15 µm is from ISOCAM (Hopwood et al. 2010) on ISO. The lower limits from MIPS at 24, 70 and 160 µm on Spitzer are provided by Bethermin et al. (2010)(solid blue) and by Chary et al. (2004), Frayer et al. (2006) and Dole et al. (2006) (solid gold, open gold and open green, respectively). Lower limits fromHerschel number counts (Berta et al. 2010) are shown as solid red triangles. In the submillimetre, limits are presented from the BLAST experiment (greenpoints; Devlin et al. 2009). Direct detection: in the optical, orange hexagons are based on data from the Pioneer 10/11 Imaging Photopolarimeter (Matsuokaet al. 2011), which are consistent with the older determination of Toller (1983). The blue star is a determination from Mattila et al. (2011), and the triangleat 520 nm is an upper limit from the same. The points at 1.25, 2.2 and 3.5 µm are based upon DIRBE data with foreground subtraction: Wright (2001, darkred squares), Cambresy et al. (2001, orange crosses), Levenson & Wright (2008, red diamond), Gorjian et al. (2000, purple open hexes), Wright & Reese(2000, green square) and Levenson et al. (2007, red asterisks). In the far-IR, direct detection measurements are shown from DIRBE (Schlegel, Finkbeiner &Davis 1998; Wright 2004, solid red circles and blue stars) and FIRAS (Fixsen et al. 1998, purple bars). Blue–violet open squares are from IR backgroundmeasurements with the AKARI satellite (Matsuura et al. 2011).

Table 1. The integrated flux of the local EBL in our models (WMAP5 with evolving and fixeddust parameters, and the C!CDM model) and the model of D11. Units are nW m!2 sr!1.

Wavelength range WMAP5 (fiducial) WMAP5+fixed C!CDM D11

Optical–near-IR peak (0.1–8 µm) 29.01 24.34 26.15 24.47Mid-IR (8–50 µm) 4.89 5.16 5.86 5.24

Far-IR peak (50–500 µm) 21.01 22.94 24.08 39.48Total (0.1–500 µm) 54.91 52.44 56.09 69.19

C" 2012 The Authors, MNRAS 422, 3189–3207Monthly Notices of the Royal Astronomical Society C" 2012 RAS

WMAP1

Propagating D+11 errors in SED fits and redshift extrapolation

The evolution of the EBL with redshift is shown graphically in Fig. 5, in two ways:

in physical and co-moving coordinates. The left panel shows that the EBL was much higher in the past, especially in the optical and near-IR and in the far-IR. The right panel shows how the present-day EBL was generated as a function of redshift. This EBL evolution must be taken into account in calculating attenuation of gamma rays from all but the nearest extragalactic sources. The change in the functional form of the EBL means that a simple z-dependent scaling model is inadequate.

FIGURE 5. The evolution of the EBL in our WMAP5 Fiducial model. This is plotted on the left panel

in standard units. The right panel shows the build-up of the present-day EBL by plotting the same quantities in comoving units. The redshifts from 0 to 2.5 are shown by the different line types in the

key in the left panel. (From Fig. 5 of [9].)

GAMMA RAY ATTENUATION

Gamma ray attenuation due to γγ → e+e− is calculated by integrating the cross section times the proper density of background photons along the line of sight to the emitting redshift, and integrating over the scattering angle θ, where θ = π corresponds to a head-on collision. The most probable scattering angle is θ ≈ π/2. If we assume θ = π/2, then the characteristic wavelength λbg of the background photons that will most strongly affect a gamma ray of energy Eγ is given by λbg = 1.2 (Eγ /TeV) µm.

We have calculated gamma-ray attenuation as a function of the redshift of the source and the observed gamma-ray energy, from the evolving EBL determined both observationally and from our SAM calculations. This is shown in the left panel of Fig. 6.

A more general way to show the EBL attenuation is to plot the “Attenuation Edge” redshift where the optical depth τ reaches a certain value as a function of gamma-ray energy, which is presented in the right panel of Fig. 6 out to redshift 5 for τ = 1, 3, and

The evolution of the EBL in our WMAP5 Fiducial model. This is plotted on the left panel in standard units. The right panel shows the build-up of the present-day EBL by plotting the same quantities in comoving units. The redshifts from 0 to 2.5 are shown by the different line types in the key in the left panel. Gilmore, Somerville, Primack, & Domínguez (2012)

Evolution of the EBL

Physical Coordinates Co-moving Coordinates

Predicted Gamma Ray Attenuation

Increasing distance causes absorption features to increase in magnitude and appear at lower energies. The plateau seen between 1 and 10 TeV at low z is a product of the mid-IR valley in the EBL spectrum.


3198 R. C. Gilmore et al.

Figure 7. The attenuation e!! of gamma-rays versus gamma-ray energy,for sources at z = 0.03, 0.1, 0.25, 0.5 and 1. Results are compared for ourfiducial WMAP5 (solid) and WMAP5+fixed (dash–dotted violet) models,as well as the model of D11 (red dash–dotted). Increasing distance causesabsorption features to increase in magnitude and appear at lower energies.The plateau seen between 1 and 10 TeV at low redshift is a product of themid-IR valley in the EBL spectrum.

Figure 8. The gamma-ray attenuation edges for the WMAP5 (solid black)and WMAP5+fixed (dash–dotted violet) models and model of D11 (reddash–dotted). The curves show the redshift at which the pair productionoptical depth ! reaches the indicated value for a particular observed gamma-ray energy. The groups of curves from lower left to upper right are thecontours for ! = 1, 3 and 10. We have included thin lines to guide the eyeat 50 and 100 GeV.

3.3 Results for TeV blazars

Today, exploration in the VHE (30 GeV to 30 TeV) regime isled by >10-m-class imaging atmospheric Cherenkov telescopes(IACTs) including the Very Energetic Radiation Imaging Tele-scope Array System (VERITAS; Maier et al. 2008), High EnergyStereoscopic System (HESS; Hinton 2004) and Major AtmosphericGamma-Ray Imaging Cherenkov (MAGIC; Cortina 2005) experi-

ments, and by the Large Area Telescope (LAT) instrument on theFermi gamma-ray space telescope (Atwood et al. 2009) and alsoAGILE (Tavani et al. 2008).

The Fermi LAT spends most of its time in an-all sky survey mode,and with its large area of view is therefore an ideal instrument forfinding high-energy sources. The 11-month source catalogue lists685 high-energy sources associated with blazar candidates (Abdoet al. 2010a). While the Fermi LAT has an energy range of 20 MeVto "300 GeV, it has a much smaller effective area than the currentgeneration of ground-based instruments, and data from the instru-ment is therefore most useful for our purposes at energies below thethreshold of these IACTs, 50–100 GeV. A detailed analysis of theEBL constraints available from all Fermi observations of blazarsand gamma-ray bursts (GRBs) to date was the subject of a recentpaper by the Fermi collaboration, Abdo et al. (2010b). Current lim-its on the EBL available from Fermi observations do not constrainthe UV flux predicted in Gilmore et al. (2009) or in the modelspresented here.

In this section and the following section, we will focus on theeffect of the optical–IR EBL on AGN-type sources by IACTs at!100 GeV. Ground-based detectors searching above 100 GeV haveidentified 37 extragalactic AGN-like sources at the time of thiswriting, including 32 BL Lac objects, radio galaxies M87 and Cen-taurus A, and the flat-spectrum radio quasars 3C 279, PKS 1510!08and PKS 1222+21. With the exception of the radio galaxies theseobjects are all blazars, accreting AGN which generate tightlybeamed relativistic jets that are oriented at a small angle relativeto our line of sight. While they account for the large majority of de-tected sources above 100 GeV, BL Lac objects are themselves onlya small subset ("20 per cent) of all blazar sources, the other 80 percent being flat spectrum radio quasars like 3C 279.

3.3.1 Constraints from gamma-ray observations

While uncertainties and likely variation in the intrinsic spectrum ofblazars make it impossible to directly link the observed spectrumto EBL attenuation, it is possible to translate limits on the spec-tra to EBL constraints. The standard assumption in placing limitson the EBL from individual spectra is that the reconstructed in-trinsic spectrum should not have a spectral index harder than 1.5,that is, " # 1.5 where dN/dE $ E!" for photon count N, or al-ternatively dF/dE $ E!("!1) for flux F. This figure comes aboutboth on the basis of experimental observations (no observed VHEspectrum is harder than this value) and theoretical arguments. Thestandard value for a single-zone synchrotron-self-Compton (SSC)spectrum is " = (# + 1)/2; here !# is the spectral index of theshock-accelerated electrons, which is not harder than 2.0 in mostacceleration models with radiative cooling (Aharonian 2001). Thiscan be invalidated by assuming a non-standard spectrum for theelectrons; a low energy cut-off in the electron energy will lead toinverse-Compton accelerated photons with an index as low as " =2/3 (Katarzynski et al. 2006).

The most recent limits on the EBL come from observations ofblazars at more distant redshifts (z > 0.1) that have been detectedby the current generation of ground-based atmospheric Cherenkovtelescopes (ACTs). Observation by HESS of two blazars at z =0.165 and 0.186 were used to set limits on the near-IR EBL basedon the " # 1.5 criterion (Aharonian et al. 2006); in this case themaximal limit was the model of Primack et al. (2001) multiplied bya factor of 0.45. Another paper by the HESS group set constraintsfrom blazar 1ES 0229+200 at z = 0.1396 (Aharonian et al. 2007b).

C% 2012 The Authors, MNRAS 422, 3189–3207Monthly Notices of the Royal Astronomical Society C% 2012 RAS

WMAP5 Fiducial

3198

R. C

. Gilm

ore

etal

.

Fig

ure

7.T

heat

tenu

atio

ne!

!of

gam

ma-

rays

vers

usga

mm

a-ra

yen

ergy

,fo

rso

urce

sat

z=

0.03

, 0.1

, 0.2

5,0.

5an

d1.

Res

ults

are

com

pare

dfo

rou

rfid

ucia

lW

MA

P5

(sol

id)

and

WM

AP

5+fix

ed(d

ash–

dotte

dvi

olet

)m

odel

s,as

wel

l as

the

mod

elof

D11

(red

dash

–dot

ted)

. Inc

reas

ing

dist

ance

caus

esab

sorp

tion

feat

ures

toin

crea

sein

mag

nitu

dean

dap

pear

atlo

wer

ener

gies

.T

hepl

atea

use

enbe

twee

n1

and

10Te

Vat

low

reds

hift

isa

prod

uct o

fth

em

id-I

Rva

lley

inth

eE

BL

spec

trum

.

Fig

ure

8.T

hega

mm

a-ra

yat

tenu

atio

ned

ges

for

the

WM

AP

5(s

olid

blac

k)an

dW

MA

P5+

fixed

(das

h–do

tted

viol

et)

mod

els

and

mod

elof

D11

(red

dash

–dot

ted)

.T

hecu

rves

show

the

reds

hift

atw

hich

the

pair

prod

uctio

nop

tical

dept

h!

reac

hes t

hein

dica

ted

valu

efo

r apa

rtic

ular

obse

rved

gam

ma-

ray

ener

gy.

The

grou

psof

curv

esfr

omlo

wer

left

toup

per

righ

tar

eth

eco

ntou

rsfo

r!

=1,

3an

d10

. We

have

incl

uded

thin

lines

togu

ide

the

eye

at50

and

100

GeV

.

3.3

Res

ults

for

TeV

blaz

ars

Toda

y,ex

plor

atio

nin

the

VH

E(3

0G

eVto

30Te

V)

regi

me

isle

dby

>10

-m-c

lass

imag

ing

atm

osph

eric

Che

renk

ovte

lesc

opes

(IA

CT

s)in

clud

ing

the

Ver

yE

nerg

etic

Rad

iatio

nIm

agin

gTe

le-

scop

eA

rray

Syst

em(V

ER

ITA

S;M

aier

etal

. 200

8), H

igh

Ene

rgy

Ster

eosc

opic

Syst

em(H

ESS

; Hin

ton

2004

) and

Maj

orA

tmos

pher

icG

amm

a-R

ayIm

agin

gC

here

nkov

(MA

GIC

; Cor

tina

2005

)ex

peri

-

men

ts, a

ndby

the

Lar

geA

rea

Tele

scop

e(L

AT

)in

stru

men

t on

the

Ferm

iga

mm

a-ra

ysp

ace

tele

scop

e(A

twoo

det

al.

2009

)an

dal

soA

GIL

E(T

avan

i et a

l.20

08).

The

Ferm

iLA

Tsp

ends

mos

t of i

tstim

ein

an-a

llsk

ysu

rvey

mod

e,an

dw

ithits

larg

ear

eaof

view

isth

eref

ore

anid

eal i

nstr

umen

t for

findi

nghi

gh-e

nerg

yso

urce

s.T

he11

-mon

thso

urce

cata

logu

elis

ts68

5hi

gh-e

nerg

yso

urce

sas

soci

ated

with

blaz

arca

ndid

ates

(Abd

oet

al. 2

010a

).W

hile

the

Ferm

i LA

Tha

san

ener

gyra

nge

of20

MeV

to"

300

GeV

, it h

asa

muc

hsm

alle

ref

fect

ive

area

than

the

curr

ent

gene

ratio

nof

grou

nd-b

ased

inst

rum

ents

, and

data

from

the

inst

ru-

men

t is

ther

efor

em

ost u

sefu

l for

our p

urpo

ses

aten

ergi

esbe

low

the

thre

shol

dof

thes

eIA

CT

s,50

–100

GeV

. Ade

taile

dan

alys

isof

the

EB

Lco

nstr

aint

sav

aila

ble

from

all

Ferm

iob

serv

atio

nsof

blaz

ars

and

gam

ma-

ray

burs

ts(G

RB

s)to

date

was

the

subj

ect o

fa

rece

ntpa

per b

yth

eFe

rmi c

olla

bora

tion,

Abd

oet

al. (

2010

b). C

urre

ntlim

-its

onth

eE

BL

avai

labl

efr

omFe

rmi o

bser

vatio

nsdo

not c

onst

rain

the

UV

flux

pred

icte

din

Gilm

ore

etal

.(2

009)

orin

the

mod

els

pres

ente

dhe

re.

Inth

isse

ctio

nan

dth

efo

llow

ing

sect

ion,

we

will

focu

son

the

effe

ctof

the

optic

al–I

RE

BL

onA

GN

-typ

eso

urce

sby

IAC

Ts

at!

100

GeV

. Gro

und-

base

dde

tect

ors

sear

chin

gab

ove

100

GeV

have

iden

tified

37ex

trag

alac

ticA

GN

-lik

eso

urce

sat

the

time

ofth

isw

ritin

g,in

clud

ing

32B

LL

acob

ject

s,ra

dio

gala

xies

M87

and

Cen

-ta

urus

A, a

ndth

efla

t-sp

ectr

umra

dio

quas

ars 3

C27

9,PK

S15

10!

08an

dPK

S12

22+

21. W

ithth

eex

cept

ion

ofth

era

dio

gala

xies

thes

eob

ject

sar

eal

lbl

azar

s,ac

cret

ing

AG

Nw

hich

gene

rate

tight

lybe

amed

rela

tivis

ticje

tsth

atar

eor

ient

edat

asm

all

angl

ere

lativ

eto

our l

ine

ofsi

ght.

Whi

leth

eyac

coun

t for

the

larg

em

ajor

ityof

de-

tect

edso

urce

sab

ove

100

GeV

, BL

Lac

obje

cts

are

them

selv

eson

lya

smal

l sub

set (

"20

per c

ent)

ofal

l bla

zar s

ourc

es, t

heot

her 8

0pe

rce

ntbe

ing

flat s

pect

rum

radi

oqu

asar

slik

e3C

279.

3.3.

1C

onst

rain

tsfr

omga

mm

a-ra

yob

serv

atio

nsW

hile

unce

rtai

ntie

san

dlik

ely

vari

atio

nin

the

intr

insi

csp

ectr

umof

blaz

ars

mak

eit

impo

ssib

leto

dire

ctly

link

the

obse

rved

spec

trum

toE

BL

atte

nuat

ion,

itis

poss

ible

totr

ansl

ate

limits

onth

esp

ec-

tra

toE

BL

cons

trai

nts.

The

stan

dard

assu

mpt

ion

inpl

acin

glim

itson

the

EB

Lfr

omin

divi

dual

spec

tra

isth

atth

ere

cons

truc

ted

in-

trin

sic

spec

trum

shou

ldno

tha

vea

spec

tral

inde

xha

rder

than

1.5,

that

is,"

#1.

5w

here

dN/d

E$

E!

"fo

rph

oton

coun

tN

,or

al-

tern

ativ

ely

dF/d

E$

E!

("!

1)fo

rflu

xF

.T

his

figur

eco

mes

abou

tbo

thon

the

basi

sof

expe

rim

enta

l obs

erva

tions

(no

obse

rved

VH

Esp

ectr

umis

hard

erth

anth

isva

lue)

and

theo

retic

alar

gum

ents

. The

stan

dard

valu

efo

ra

sing

le-z

one

sync

hrot

ron-

self

-Com

pton

(SSC

)sp

ectr

umis

"=

(#+

1)/2

;he

re!

#is

the

spec

tral

inde

xof

the

shoc

k-ac

cele

rate

del

ectr

ons,

whi

chis

not

hard

erth

an2.

0in

mos

tac

cele

ratio

nm

odel

sw

ithra

diat

ive

cool

ing

(Aha

roni

an20

01).

Thi

sca

nbe

inva

lidat

edby

assu

min

ga

non-

stan

dard

spec

trum

for

the

elec

tron

s;a

low

ener

gycu

t-of

fin

the

elec

tron

ener

gyw

illle

adto

inve

rse-

Com

pton

acce

lera

ted

phot

ons

with

anin

dex

aslo

was

"=

2/3

(Kat

arzy

nski

etal

. 200

6).

The

mos

tre

cent

limits

onth

eE

BL

com

efr

omob

serv

atio

nsof

blaz

ars

atm

ore

dist

ant r

edsh

ifts

(z>

0.1)

that

have

been

dete

cted

byth

ecu

rren

t gen

erat

ion

ofgr

ound

-bas

edat

mos

pher

icC

here

nkov

tele

scop

es(A

CT

s).

Obs

erva

tion

byH

ESS

oftw

obl

azar

sat

z=

0.16

5an

d0.

186

wer

eus

edto

set l

imits

onth

ene

ar-I

RE

BL

base

don

the

"#

1.5

crite

rion

(Aha

roni

anet

al. 2

006)

;in

this

case

the

max

imal

limit

was

the

mod

elof

Prim

ack

etal

. (20

01) m

ultip

lied

bya

fact

orof

0.45

. Ano

ther

pape

rby

the

HE

SSgr

oup

set c

onst

rain

tsfr

ombl

azar

1ES

0229

+20

0at

z=

0.13

96(A

haro

nian

etal

. 200

7b).

C%20

12T

heA

utho

rs, M

NR

AS

422,

3189

–320

7M

onth

lyN

otic

esof

the

Roy

alA

stro

nom

ical

Soci

ety

C%20

12R

AS

WMAP5 Fixed Domiínguez+11


Figure 4. The predicted z = 0 EBL spectrum from our fiducial WMAP5 model (solid black) and WMAP5+fixed (dash–dotted violet) dust parameters, andC!CDM (dotted black) models, compared with experimental constraints at a number of wavelengths. D11 is shown for comparison in dashed–dotted red withthe shaded area indicating the uncertainty region. Data: upward pointing arrows indicate lower bounds from number counts; other symbols are results fromdirect detection experiments. Note that some points have been shifted slightly in wavelength for clarity. Lower limits: the blue–violet triangles are results fromHST and Space Telescope Imaging Spectrograph (STIS; Gardner et al. 2000), while the purple open triangles are from GALEX (Xu et al. 2005). The solid greenand red triangles are from the Hubble Deep Field (Madau & Pozzetti 2000) and Ultra Deep Field (Dolch & Ferguson, in preparation), respectively, combinedwith ground-based data, and the solid purple triangle is from a measurement by the Large Binocular Camera (Grazian et al. 2009). In the near-IR J, H and Kbands, open violet points are the limits from Keenan et al. (2010). Open red triangles are from IRAC on Spitzer (Fazio et al. 2004), and the purple triangle at15 µm is from ISOCAM (Hopwood et al. 2010) on ISO. The lower limits from MIPS at 24, 70 and 160 µm on Spitzer are provided by Bethermin et al. (2010)(solid blue) and by Chary et al. (2004), Frayer et al. (2006) and Dole et al. (2006) (solid gold, open gold and open green, respectively). Lower limits fromHerschel number counts (Berta et al. 2010) are shown as solid red triangles. In the submillimetre, limits are presented from the BLAST experiment (greenpoints; Devlin et al. 2009). Direct detection: in the optical, orange hexagons are based on data from the Pioneer 10/11 Imaging Photopolarimeter (Matsuokaet al. 2011), which are consistent with the older determination of Toller (1983). The blue star is a determination from Mattila et al. (2011), and the triangleat 520 nm is an upper limit from the same. The points at 1.25, 2.2 and 3.5 µm are based upon DIRBE data with foreground subtraction: Wright (2001, darkred squares), Cambresy et al. (2001, orange crosses), Levenson & Wright (2008, red diamond), Gorjian et al. (2000, purple open hexes), Wright & Reese(2000, green square) and Levenson et al. (2007, red asterisks). In the far-IR, direct detection measurements are shown from DIRBE (Schlegel, Finkbeiner &Davis 1998; Wright 2004, solid red circles and blue stars) and FIRAS (Fixsen et al. 1998, purple bars). Blue–violet open squares are from IR backgroundmeasurements with the AKARI satellite (Matsuura et al. 2011).

Table 1. The integrated flux of the local EBL in our models (WMAP5 with evolving and fixeddust parameters, and the C!CDM model) and the model of D11. Units are nW m!2 sr!1.

Wavelength range WMAP5 (fiducial) WMAP5+fixed C!CDM D11

Optical–near-IR peak (0.1–8 µm) 29.01 24.34 26.15 24.47Mid-IR (8–50 µm) 4.89 5.16 5.86 5.24

Far-IR peak (50–500 µm) 21.01 22.94 24.08 39.48Total (0.1–500 µm) 54.91 52.44 56.09 69.19

C" 2012 The Authors, MNRAS 422, 3189–3207Monthly Notices of the Royal Astronomical Society C" 2012 RAS

Gamma Ray Attenuation due to γγ → e+e-

If we know the intrinsic spectrum, we can infer the optical depth τ(E,z) from the observed spectrum. In practice, we typically assume that dN/dE|int is not harder than E-Γ with Γ = 1.5, since local sources have Γ ≥ 2. More conservatively, we can assume that Γ ≥ 2/3.

Illustration: Mazin & Raue

Reconstructed Blazar Spectral Indexes

Γ=1.5

With our SAM based on current WMAP5 cosmological parameters and Spitzer (Rieke+09) dust emission templates, all high redshift blazars have spectral indexes Γ≥1.5, as expected from nearby sources.

(Of course, Felix can make them much harder!)

1ES 0229+200H 1426+428


Cosmic Gamma-Ray Horizon


100 GeVThreshold

50 GeVThreshold

With a 50 GeV threshold, we see to z≈1.5-3with less than 1/e attenuation!

3198 R. C. Gilmore et al.

Figure 7. The attenuation e!! of gamma-rays versus gamma-ray energy,for sources at z = 0.03, 0.1, 0.25, 0.5 and 1. Results are compared for ourfiducial WMAP5 (solid) and WMAP5+fixed (dash–dotted violet) models,as well as the model of D11 (red dash–dotted). Increasing distance causesabsorption features to increase in magnitude and appear at lower energies.The plateau seen between 1 and 10 TeV at low redshift is a product of themid-IR valley in the EBL spectrum.

Figure 8. The gamma-ray attenuation edges for the WMAP5 (solid black)and WMAP5+fixed (dash–dotted violet) models and model of D11 (reddash–dotted). The curves show the redshift at which the pair productionoptical depth ! reaches the indicated value for a particular observed gamma-ray energy. The groups of curves from lower left to upper right are thecontours for ! = 1, 3 and 10. We have included thin lines to guide the eyeat 50 and 100 GeV.

3.3 Results for TeV blazars

Today, exploration in the VHE (30 GeV to 30 TeV) regime isled by >10-m-class imaging atmospheric Cherenkov telescopes(IACTs) including the Very Energetic Radiation Imaging Tele-scope Array System (VERITAS; Maier et al. 2008), High EnergyStereoscopic System (HESS; Hinton 2004) and Major AtmosphericGamma-Ray Imaging Cherenkov (MAGIC; Cortina 2005) experi-

ments, and by the Large Area Telescope (LAT) instrument on theFermi gamma-ray space telescope (Atwood et al. 2009) and alsoAGILE (Tavani et al. 2008).

The Fermi LAT spends most of its time in an-all sky survey mode,and with its large area of view is therefore an ideal instrument forfinding high-energy sources. The 11-month source catalogue lists685 high-energy sources associated with blazar candidates (Abdoet al. 2010a). While the Fermi LAT has an energy range of 20 MeVto "300 GeV, it has a much smaller effective area than the currentgeneration of ground-based instruments, and data from the instru-ment is therefore most useful for our purposes at energies below thethreshold of these IACTs, 50–100 GeV. A detailed analysis of theEBL constraints available from all Fermi observations of blazarsand gamma-ray bursts (GRBs) to date was the subject of a recentpaper by the Fermi collaboration, Abdo et al. (2010b). Current lim-its on the EBL available from Fermi observations do not constrainthe UV flux predicted in Gilmore et al. (2009) or in the modelspresented here.

In this section and the following section, we will focus on theeffect of the optical–IR EBL on AGN-type sources by IACTs at!100 GeV. Ground-based detectors searching above 100 GeV haveidentified 37 extragalactic AGN-like sources at the time of thiswriting, including 32 BL Lac objects, radio galaxies M87 and Cen-taurus A, and the flat-spectrum radio quasars 3C 279, PKS 1510!08and PKS 1222+21. With the exception of the radio galaxies theseobjects are all blazars, accreting AGN which generate tightlybeamed relativistic jets that are oriented at a small angle relativeto our line of sight. While they account for the large majority of de-tected sources above 100 GeV, BL Lac objects are themselves onlya small subset ("20 per cent) of all blazar sources, the other 80 percent being flat spectrum radio quasars like 3C 279.

3.3.1 Constraints from gamma-ray observations

While uncertainties and likely variation in the intrinsic spectrum ofblazars make it impossible to directly link the observed spectrumto EBL attenuation, it is possible to translate limits on the spec-tra to EBL constraints. The standard assumption in placing limitson the EBL from individual spectra is that the reconstructed in-trinsic spectrum should not have a spectral index harder than 1.5,that is, " # 1.5 where dN/dE $ E!" for photon count N, or al-ternatively dF/dE $ E!("!1) for flux F. This figure comes aboutboth on the basis of experimental observations (no observed VHEspectrum is harder than this value) and theoretical arguments. Thestandard value for a single-zone synchrotron-self-Compton (SSC)spectrum is " = (# + 1)/2; here !# is the spectral index of theshock-accelerated electrons, which is not harder than 2.0 in mostacceleration models with radiative cooling (Aharonian 2001). Thiscan be invalidated by assuming a non-standard spectrum for theelectrons; a low energy cut-off in the electron energy will lead toinverse-Compton accelerated photons with an index as low as " =2/3 (Katarzynski et al. 2006).

The most recent limits on the EBL come from observations ofblazars at more distant redshifts (z > 0.1) that have been detectedby the current generation of ground-based atmospheric Cherenkovtelescopes (ACTs). Observation by HESS of two blazars at z =0.165 and 0.186 were used to set limits on the near-IR EBL basedon the " # 1.5 criterion (Aharonian et al. 2006); in this case themaximal limit was the model of Primack et al. (2001) multiplied bya factor of 0.45. Another paper by the HESS group set constraintsfrom blazar 1ES 0229+200 at z = 0.1396 (Aharonian et al. 2007b).

C% 2012 The Authors, MNRAS 422, 3189–3207Monthly Notices of the Royal Astronomical Society C% 2012 RAS

Mon. Not. R. Astron. Soc. 000, 1–?? (2011) Printed 31 October 2012 (MN LATEX style file v2.2)

Detection of the cosmic �-ray horizon from

multiwavelength observations of blazars

A. Domınguez

1?, J. D. Finke

2, F. Prada

3,4,5, J. R. Primack

6, F. S. Kitaura

7, B. Siana

1,

S. Ciprini

8,9, J. Knodlseder

10,11and D. Paneque

12,131Dept. Physics & Astronomy, University of California, Riverside, CA 92521, USA2U.S. Naval Research Laboratory, Space Science Division, Code 7653, 4555 Overlook Ave. SW, Washington, DC, 20375 USA3Campus of International Excellence UAM+CSIC, Cantoblanco, E-28049 Madrid, Spain4Instituto de Fısica Teorica, (UAM/CSIC), Universidad Autonoma de Madrid, Cantoblanco, E-28049 Madrid, Spain5Instituto de Astrofısica de Andalucıa (CSIC), Glorieta de la Astronomıa, E-18080 Granada, Spain6Department of Physics, University of California, Santa Cruz, CA 95064, USA7Leibniz-Institut fur Astrophysik (AIP), An der Sternwarte 16, D-14482 Potsdam, Germany8INAF, Observatory of Rome, Monte Porzio Catone, Roma, Italy9Agenzia Spaziale Italiana (ASI) Science Data Center, I-00044 Frascati (Roma), Italy10Institut de Recherche en Astrophysique et Planetologie (CESR), F-31028 Toulouse Cedex 4, France11GAHEC, Universite de Toulouse, UPS-OMP, IRAP, Toulouse, France12Kavli Institute for Particle Astrophysics and Cosmology, SLAC, Stanford University, Stanford, CA 94305, USA13Max-Planck-Institut fur Physik, D-80805 Munchen, Germany

31 October 2012

ABSTRACTThe first statistically significant detection of the cosmic �-ray horizon (CGRH) thatis independent of any extragalactic background light (EBL) model is presented. TheCGRH is a fundamental quantity in cosmology. It gives an estimate of the opacityof the Universe to very-high energy (VHE) �-ray photons due to photon-photon pairproduction with the EBL. The only estimations of the CGRH to date are predictionsfrom EBL models and lower limits from �-ray observations of cosmological blazars and�-ray bursts. Here, we present synchrotron/synchrotron self-Compton models (SSC) ofthe spectral energy distributions of 15 blazars based on (almost) simultaneous observa-tions from radio up to the highest energy �-rays taken with the Fermi satellite. Thesesynchrotron/SSC models predict the unattenuated VHE fluxes, which are comparedwith the observations by imaging atmospheric Cherenkov telescopes. This comparisonprovides an estimate of the optical depth of the EBL, which allows a derivation of theCGRH through a maximum likelihood analysis that is EBL-model independent. Wefind that the observed CGRH is compatible with the current knowledge of the EBL.

Key words: cosmology: observations - di↵use radiation – galaxies: formation – galax-ies: evolution – gamma-rays: observations

1 INTRODUCTION

Very-high energy (VHE; 30 GeV–30 TeV) photons do nottravel unimpeded through cosmological distances in the Uni-verse. A flux attenuation is expected due to photon-photonpair production with the lower energy photons of the extra-galactic background light (EBL) in the ultraviolet, optical,and infrared (IR) (e.g., Nikishov 1962; Gould & Schreder1967; Stecker, de Jager & Salamon 1992; Salamon & Stecker1998). The EBL is the radiation emitted by star formationprocesses (star light and star light absorbed/re-emitted by

? E-mail: [email protected]

dust) plus a small contribution from active galactic nuclei(AGNs) integrated over redshift over all the cosmic star-formation history (e.g., Hauser & Dwek 2001). Due to theproperties of the interaction, a VHE photon of a given en-ergy interacts mainly with an EBL photon within a welldefined and narrow wavelength range. Therefore, a signa-ture of the EBL spectral distribution is expected in the ob-served VHE spectra of extragalactic sources (Ackermann etal. 2012b).

An interesting feature in the observed VHE spectra ofextragalactic sources as a consequence of EBL absorption isgiven by the cosmic �-ray horizon (CGRH), which has notbeen clearly observed yet. The CGRH is by definition theenergy at which the optical depth of the photon-photon pair

c� 2011 RAS




Collaborators:




Presented at the 4th Fermi Symposium in Monterey, CA

Approved by the Fermi Publication Board 31 Oct 2012Submitted but not on arXiv

SED multiwavelength fitsSED multiwavelength fits

A one-zone synchrotron/SSC model is fit to the multiwavelength data excluding the

Cherenkov data, which are EBL attenuated. Then, this fit is extrapolated to the VHE regime

representing the intrinsic VHE spectrum. Technique similar to Mankuzhiyil et al. 2010.

PKS 2155-304

z = 0.116

Variability time

scale 104 s (fast)

Preliminary

Preliminary 1ES 1218+304

z = 0.182

Variability time

scale 105 s (slow)

Preliminary

Preliminary

Domínguez+12

Cosmic Cosmic γ-ray Horizon: resultsγ-ray Horizon: results

There are 4 out of 15 cases where our maximum likelihood methodology could not be applied since the prediction from the There are 4 out of 15 cases where our maximum likelihood methodology could not be applied since the prediction from the

synchrotron/SSC model was lower than the detected flux by the Cherenkov telescopes.synchrotron/SSC model was lower than the detected flux by the Cherenkov telescopes.

Two other cases where the statistical uncertainties were too high to set any constraint on E0.Two other cases where the statistical uncertainties were too high to set any constraint on E0.

Propagating D+11 errors in SED fits and redshift extrapolation

Domínguez+12

Presented at the 4th Fermi Symposium in Monterey, CA

Published Online in Science 4 Nov 2012

The Imprint of the EBL in

the Spectra of Blazars

Marco Ajello1,2, Anita Reimer3, Rolf Buehler1 on behalf of the Fermi-LAT

collaboration

1KIPAC, 2UCB, 3Univ. Innsbruck

1 Ackermann+12, in ScienceExpress Today!

The Imprint of the EBL in

the Spectra of Blazars

Marco Ajello1,2, Anita Reimer3, Rolf Buehler1 on behalf of the Fermi-LAT

collaboration

1KIPAC, 2UCB, 3Univ. Innsbruck

1 Ackermann+12, in ScienceExpress Today!

Reports

/ http://www.sciencemag.org/content/early/recent / 1 November 2012 / Page 1/ 10.1126/science.1227160

The bulk of the intergalactic gas in the universe must have been reionized between the epoch of cosmic recombination, when the uni-verse was only 300,000 years old (z 1100 ), and 1 billion years later (z as indicated observationally by the spectra of distant quasi-stellar (6 objects (1). However, the sources, modes and nature of this cosmic reionization are largely unknown because most of this redshift range has yet to be explored. Photoionization by UV radiation, produced by the first stars and galaxies of the universe, represents the primary suspect for the ionizing process (2, 3). Direct detection of the UV radiation fields is thus of fundamental importance, but at present extremely difficult (3).

An indirect but powerful means of probing the diffuse radiation ILHOGV�LV�WKURXJK�Ȗ-Ȗ�DEVRUSWLRQ�RI�KLJK-energy gamma rays (4–6). In this

process, a gamma-ray photon of ener-gy EȖ and an EBL photon of energy EEBL annihilate and create an electron-positron pair. This process occurs for head-on collisions when (e.g.) EȖ ×EEBL �� me c2)2, where me c2 is the rest mass energy of the electron. This introduces an attenuation in the spectra of gamma-ray sources above a critical gamma-ray energy of Ecrit(z�� §�170(1+z)í�� GeV (7, 8).

The detection of the gamma-ray horizon (i.e., the point beyond which the emission of gamma-ray sources is strongly attenuated) is one of the pri-mary scientific drivers of the Fermi Gamma-ray Space Telescope (9–11). Several attempts have been made in the past but none detected the long-sought EBL attenuation (12–14). So far, limits on the EBL density have been inferred from the absence of ab-sorption features in the spectra of indi-vidual blazars (13, 15), distant galaxies with bright gamma-ray emission pow-ered by matter accreting onto central, massive black holes. While this feature is indeed difficult to constrain for a single source, we show that it is de-tected collectively in the gamma-ray spectra of a sample of blazars as a cut-off that changes amplitude and energy with redshift. We searched for an at-tenuation of the spectra of blazars in the 1-500 GeV band using the first 46 months of observations of the Large Area Telescope (LAT) on board the Fermi satellite. At these energies gamma rays are absorbed by EBL photons in the optical to UV range. Thanks to the large energy and redshift coverage, Fermi-LAT measures the intrinsic (i.e., unabsorbed) spectrum up to 100 GeV for any blazar at z < 0.2, and up to 15 GeV for any redshift.

The LAT has detected > 1000 blazars to date (16). We restricted our search to a subset of 150 blazars of the BL Lacertae (BL Lac) type that are significantly detected above 3 GeV, because of the expected lack of intrin-

sic absorption (17). The sample covers a redshift range 0.03-1.6 (18, 19). 7KH�FULWLFDO�HQHUJ\�LV�WKHUHIRUH�DOZD\V��*H9��ZKLFK�PHDQV�WKDW�WKH�spectrum measured below this energy is unabsorbed and a true represen-tation of the intrinsic spectrum of the source. We thus determined the intrinsic source spectrum relying on data between 1 GeV and the critical energy Ecrit and extrapolated it to higher energies. By combining all the spectra we were able to determine, the average deviation, above the crit-ical energy, of the measured spectra from the intrinsic ones, which ulti-PDWHO\�SURYLGHV�D�PHDVXUHPHQW�RI�WKH�RSWLFDO�GHSWK�ĲȖȖ.

The analysis was performed using the Fermi Science Tools (20). We determined the spectral parameters of each blazar by maximizing the likelihood of a given source model. The model comprised the galactic

The Imprint of the Extragalactic Background Light in the Gamma-Ray Spectra of Blazars M. Ackermann,1 M. Ajello,2,3* A. Allafort,2 P. Schady,4 L. Baldini,5 J. Ballet,6 G. Barbiellini,7,8 D. Bastieri,9,10 R. Bellazzini,11 R. D. Blandford,2 E. D. Bloom,2 A. W. Borgland,2 E. Bottacini,2 A. Bouvier,12 J. Bregeon,11 M. Brigida,13,14 P. Bruel,15 R. Buehler,2* S. Buson,9,10 G. A. Caliandro,16 R. A. Cameron,2 P. A. Caraveo,17 E. Cavazzuti,18 C. Cecchi,19,20 E. Charles,2 R. C. G. Chaves,6 A. Chekhtman,21 C. C. Cheung,22 J. Chiang,2 G. Chiaro,23 S. Ciprini,24,20 R. Claus,2 J. Cohen-Tanugi,25 J. Conrad,26,27,28 S. Cutini,18 F. D’Ammando,19,29,30 F. de Palma,13,14 C. D. Dermer,31 S. W. Digel,2 E. do Couto e Silva,2 A. Domínguez,12 P. S. Drell,2 A. Drlica-Wagner,2 C. Favuzzi,13,14 S. J. Fegan,15 W. B. Focke,2 A. Franckowiak,2 Y. Fukazawa,32 S. Funk,2 P. Fusco,13,14 F. Gargano,14 D. Gasparrini,18 N. Gehrels,33 S. Germani,19,20 N. Giglietto,13,14 F. Giordano,13,14 M. Giroletti,34 T. Glanzman,2 G. Godfrey,2 I. A. Grenier,6 J. E. Grove,31 S. Guiriec,33 M. Gustafsson,9 D. Hadasch,16 M. Hayashida,2,35 E. Hays,33 M. S. Jackson,36,27 T. Jogler,2 J. Kataoka,37 J. Knödlseder,38,39 M. Kuss,11 J. Lande,2 S. Larsson,26,27,40 L. Latronico,41 F. Longo,7,8 F. Loparco,13,14 M. N. Lovellette,31 P. Lubrano,19,20 M. N. Mazziotta,14 J. E. McEnery,33,42 J. Mehault,25 P. F. Michelson,2 T. Mizuno,43 C. Monte,13,14 M. E. Monzani,2 A. Morselli,44 I. V. Moskalenko,2 S. Murgia,2 A. Tramacere,45 E. Nuss,25 J. Greiner,4 M. Ohno,46 T. Ohsugi,43 N. Omodei,2 M. Orienti,34 E. Orlando,2 J. F. Ormes,47 D. Paneque,48,2 J. S. Perkins,33,49,50,51 M. Pesce-Rollins,11 F. Piron,25 G. Pivato,10 T. A. Porter,2,�2 S. Rainò,13,14 R. Rando,9,10 M. Razzano,11,12 S. Razzaque,21 A. Reimer,52,2* O. Reimer,52,2 L. C. Reyes,53 S. Ritz,12 A. Rau,4 C. Romoli,10 M. Roth,54 M. Sánchez-Conde,2 D.A. Sanchez,55 J. D. Scargle,56 C. Sgrò,11 E. J. Siskind,57 G. Spandre,11 P. Spinelli,13,14 àukasz Stawarz,46,58 D. J. Suson,59 H. Takahashi,32 T. Tanaka,2 J. G. Thayer,2 D. J. Thompson,33 L. Tibaldo,9,10 M. Tinivella,11 D. F. Torres,16,60 G. Tosti,19,20 E. Troja,33,61 T. L. Usher,2 J. Vandenbroucke,2 V. Vasileiou,25 G. Vianello,2,62 V. Vitale,44,63 A. P. Waite,2 B. L. Winer,64 K. S. Wood,31 M. Wood2

*To whom correspondence should be addressed. E-mail: [email protected] (M.A.); [email protected] (R.B.); [email protected] (A.R.)

Affiliations are listed at the end of the paper.

The light emitted by stars and accreting compact objects through the history of the universe is encoded in the intensity of the extragalactic background light (EBL). Knowledge of the EBL is important to understand the nature of star formation and galaxy evolution, but direct measurements of the EBL are limited by galactic and other foreground emissions. Here, we report an absorption feature seen in the combined spectra of a sample of gamma-ray blazars out to a redshift of z 1.6 . This feature is caused by attenuation of gamma rays by the EBL at optical to ultraviolet frequencies and allowed us to measure the EBL flux density in this frequency band.

on

Nov

embe

r 4, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

Here, we report an absorption feature seen in the combined spectra of a sample of gamma-ray blazars out to a redshift of z ∼ 1.6. This feature is caused by attenuation of gamma rays by the EBL at optical to ultraviolet frequencies and allowed us to measure the EBL flux density in this frequency band.

ABSTRACT The light emitted by stars and accreting compact objects through the history of the universe is encoded in the intensity of the extragalactic background light (EBL). Knowledge of the EBL is important to understand the nature of star formation and galaxy evolution, but direct measurements of the EBL are limited by galactic and other foreground emissions.

Analysis Procedure

•  We use 46months of P7V6 1-500 GeV data

•  We define 3 redshift bins with 50 sources each: –  z= 0-0.2, 0.2-0.5, 0.5 -1.6

•  All BL Lacs are modeled with a LogParabola spectrum

•  We perform a combined fit where: –  The spectra of all sources are fit

independently –  The spectra of all sources are modified

by a common e-b τ(E,z) term

•  We evaluate 2 cases: 1.  Null hypothesis b=0 : there is no EBL 2.  Null hypothesis b=1 : the model

predictions are correct 7

€

F(E)absorbed = F(E)int rinsic ⋅ e−b⋅τmod el

Simulated SEDs Simulated data

Fit to ‘unabsorbed’ data preliminary

We look for the collective deviation of the spectra of blazars from their intrinsic spectra

Measurement of Tau with Energy and Redshift

•  We use the composite likelihood in small energy bins to measure the collective deviation of the observed spectra from the intrinsic ones

•  The cut-off moves in z and energy as expected for EBL absorption (for low opacity models)

•  It is difficult to explain this attenuation with an intrinsic property of BL Lacs 1.  BL Lacs required to evolve across the

z=0.2 barrier 2.  Attenuation change with energy and

redshift cannot be explained by an intrinsic cut-off that changes from source to source because of redshift and blazar sequence effects

12

Best-fit EBL model!

Best-fit intrinsic cut-off!

Ackermann+12

Composite Likelihood Results: 2

•  A significant steepening in the blazars’ spectra is detected •  This is consistent with that expected by a ‘minimal’ EBL:

–  i.e. EBL at the level of galaxy counts –  4 models rejected above 3sigma

•  All the non-rejected models yield a significance of detection of 5.6-5.9 σ

•  The level of EBL is 3-4 times lower than our previous UL (Abdo+10, ApJ 723, 1082)

10

EBL Detection Significance

Model Rejection Significance

Ackermann+12

1ES0806+524

Mkn501

1ES1101-232

Mkn421

1ES1218+304

PKS2005-489

3C66A

PG1553+113PKS2155-304

10�2 10�1 100

Redshift

10�1

100

101

Cos

mic

�-r

ayho

rizon

[TeV

]

PreliminaryDomınguez+ 11

Cosmic Gamma-Ray HorizonDominguez+12 + Ackermann+12

Mon. Not. R. Astron. Soc. 420, 800–809 (2012) doi:10.1111/j.1365-2966.2011.20092.x

Constraining the near-infrared background light from Population IIIstars using high-redshift gamma-ray sources

Rudy C. Gilmore!SISSA, via Bonomea 265, 34136 Trieste, Italy

Accepted 2011 October 27. Received 2011 October 27; in original form 2011 September 1

ABSTRACTThe Fermi satellite has detected GeV emission from a number of gamma-ray bursts and activegalactic nuclei at high redshift, z ! 1.5. We examine the constraints that the detections ofgamma-rays from several of these sources place on the contribution of Population III stars to theextragalactic background light. Emission from these primordial stars, particularly redshiftedLyman " emission, can interact with gamma-rays to produce electron–positron pairs and createan optical depth to the propagation of gamma-ray emission, and the detection of emission at>10 GeV can therefore constrain the production of this background. We consider two initialmass functions for the early stars and use derived spectral energy distributions for each toput upper limits on the star formation rate density of massive early stars from redshifts 6 to10. Our limits are complementary to those set on a high near-infrared background flux byground-based TeV-scale observations and show that current data can limit star formation inthe late stages of re-ionization to less than 0.5 M! yr"1 Mpc"3. Our results also show thatthe total background flux from Population III stars must be considerably less than that fromresolved galaxies at wavelengths below 1.5 µm.

Key words: gamma-ray burst: general – stars: Population III – diffuse radiation.

1 IN T RO D U C T I O N

The re-ionization of the Universe, which took place around a red-shift of 10.5 (Komatsu et al. 2011), is generally believed to bedriven primarily by ionizing photons from early ‘Population III’(pop-III) stars. As these stars form from primordial unenriched,metal-free hydrogen and helium, they undergo a formation processthat is substantially different from that of later Population I and IIstars. Simulations of the production of pop-III stars (Abel, Bryan& Norman 2000; Bromm, Coppi & Larson 2002; Tan & McKee2004; Yoshida et al. 2006; Norman 2008) generally find an initialmass function (IMF) that is heavily biased towards high masses,10–1000 M!. A considerable fraction of the radiant energy fromthese stars is released at ionizing wavelengths (<912 Å), which al-lows re-ionization of the universe to be completed on the time-scalerequired by Lyman " forest data (Becker et al. 2001). Despite theirimportance in cosmology and impact on IGM evolution, pop-IIIstars continue to evade direct detection. Detecting the redshiftedultraviolet (UV) emission from pop-III stars is a primary goal of theupcoming James Webb Space Telescope, though even with the stateof the art sensitivity of this instrument detecting individual metal-free stars will be challenging (Rydberg, Zackrisson & Scott 2011).Searching for indirect evidence of these stars and their integrated

!E-mail: [email protected]

cosmological impact is therefore the primary way of understandingthe properties of the re-ionization-era universe.

Photon production from the re-ionization era is encoded in theevolving spectral energy distribution (SED) of the accumulatedphoton background, which we observe locally as the extragalacticbackground light (EBL). Redshifted UV radiation from pop-III starscan be expected to appear as a distinct component of the near-infrared (near-IR) portion of the EBL, and the spectral details ofthis observable light could in principle inform an observer about theredshift of re-ionization and the nature of the sources responsible(Cooray & Yoshida 2004; Kashlinsky et al. 2004; Madau & Silk2005; Fernandez & Komatsu 2006). However, observations of theabsolute intensity of the EBL in the near-IR are severely hindered bythe presence of bright galactic foregrounds, which are produced bydiverse sources including stars and the interstellar medium (ISM)of the Milky Way and ‘zodiacal light’ from dust within our solarsystem (Hauser & Dwek 2001).

It has been proposed that a high level of infrared (IR) backgroundcould be due to radiation from the first generation of stars. Multipleexperiments, most notably the DIRBE experiment on the COBEsatellite, have attempted to measure the sky brightness at near-IRwavelengths, and foreground subtraction analyses have been pre-sented by a number of authors (Gorjian, Wright & Chary 2000;Wright & Reese 2000; Cambresy, Reach, Beichman, & Jarrett 2001;Wright 2001; Levenson, Wright, & Johnson 2007), with extragalac-tic components generally exceeding the contribution of resolvable

C# 2011 The AuthorMonthly Notices of the Royal Astronomical Society C# 2011 RAS

420, 800–809 (2012) Constraining the near-infrared background light from Population III stars using high-redshift gamma-ray sourcesRudy C. Gilmore

ABSTRACT The Fermi satellite has detected GeV emission from a number of gamma-ray bursts and active galactic nuclei at high redshift, z ≳1.5. We examine the constraints that the detections of gamma-rays from several of these sources place on the contribution of Population III stars to the extragalactic background light. Emission from these primordial stars, particularly redshifted Lyman α emission, can interact with gamma-rays to produce electron–positron pairs and create an optical depth to the propagation of gamma-ray emission, and the detection of emission at >10 GeV can therefore constrain the production of this background. We consider two initial mass functions for the early stars and use derived spectral energy distributions for each to put upper limits on the star formation rate density of massive early stars from redshifts 6 to 10. Our limits are complementary to those set on a high near-infrared background flux by ground-based TeV-scale observations and show that current data can limit star formation in the late stages of re-ionization to less than 0.5 M⊙ yr−1 Mpc−3 . Our results also show that the total background flux from Population III stars must be considerably less than that from resolved galaxies at wavelengths below 1.5 μm.

808 R. C. Gilmore

Figure 6. As in the previous figure, but for a cut-off redshift zr = 9.

could be derived from future detections of high-redshift sourceswith the Fermi LAT or future telescopes. In these plots, the axesrefer to the redshift and highest observed photon energy E! of ahypothetical gamma-ray source. The source is then assumed to havea normalization at lower energy such that the expected number ofphoton counts at and above E! is 1 [Nx(>Ehigh) = 1] in the absenceof any background field. The spectrum of the source is set here to!2.25, near the mean of the sources in Table 1, and the p-EBL isignored. Given these parameters, the contours on the plots show

Figure 7. Plot of the upper bounds on the SFRD in two possible scenarioswith future Fermi GRBs, in the Larson IMF case. The solid lines showthe limits from a GRB with the same redshift and spectral characteristicsof GRB 080916C, but with a highest energy observed photon of 30 GeV(160 GeV as emitted) instead of 13.2 GeV, in combination with the sourcespreviously discussed. The dotted lines show a case with a GRB at z = 7and a highest energy observed photon at 15 GeV (120 GeV emitted). Linecolours are as in Fig. 3.

the source redshift and E! that would be required to place a givenSFRD limit on pop-III star formation at redshifts zr = 6 and 9, with2" significance. These contours are for limits derived based on asingle source; combined limits for multiple sources like those insection 3.2.2, if available, would be somewhat stronger. In Fig. 7,limits based on two hypothetical high-redshift GRBs are combinedwith the other sources of Table 1. This plot shows that new GeVsources, either at higher redshift than GRB 080916C or at a similarredshift with higher energy emission, could strongly limit a pop-IIIcontribution to star formation in the late re-ionization period.

If the Fermi satellite remains in operation for its stated lifetimegoal of 10 years from its launch date, then its mission is currentlyless than one-third complete, and we can reasonably hope to see newGRB events or high-energy AGN photons that will strengthen ourresults. The Cherenkov Telescope Array (CTA; The CTA Consor-tium 2010) is another possible source of constraining events. TheCTA will have a lower threshold energy than current-generationground-based instruments and may be able to detect sources atmuch higher redshift than currently achieved from the ground. De-tections with either of these instruments could potentially shed newlight on star formation in the re-ionization era.

AC K NOWLEDGMENTS

RCG was supported during this work by a SISSA postdoctoralfellowship, and thanks W. B. Atwood, J. Primack and A. Bouvierfor helpful discussions related to this project, and J. Colucci and theanonymous referee for reading the manuscript and providing usefulcomments. Some calculations in the paper were performed on theSISSA High-Performance Computing Cluster.

R E F E R E N C E S

Abdo A. A. et al., 2009, Sci, 323, 1688Abdo A. A. et al., 2010a, ApJ, 710, 810Abdo A. A. et al., 2010b, ApJ, 723, 1082Abel T., Bryan G. L., Norman M. L., 2000, ApJ, 540, 39Aguirre A., Schaye J., Theuns T., 2002, ApJ, 576, 1Aharonian F. et al., 2006, Nat, 440, 1018Albert J. et al., 2008, Sci, 320, 1752Atwood W. B. et al., 2009, ApJ, 697, 1071Becker R. H. et al., 2001, AJ, 122, 2850Beckman J., Giammanco C., 2010, Mem. Soc. Astron. Ital., 81, 460Bouche N., Lehnert M. D., Aguirre A., Peroux C., Bergeron J., 2007, MN-

RAS, 378, 525Bouwens R. J., Illingworth G. D., Franx M., Ford H., 2007, ApJ, 670,

928Breit G., Wheeler J. A., 1934, Phys. Rev., 46, 1087Bromm V., Coppi P. S., Larson R. B., 2002, ApJ, 564, 23Bromm V., Loeb A., 2004, New Astron., 9, 353Brown R. L., Mathews W. G., 1970, ApJ, 160, 939Cambresy L., Reach W. T., Beichman C. A., Jarrett T. H., 2001, ApJ, 555,

563Chabrier G., 2003, PASP, 115, 763Cooray A., Yoshida N., 2004, MNRAS, 351, L71Domınguez A. et al., 2011, MNRAS, 410, 2556Dwek E., Krennrich F., 2005, ApJ, 618, 657Fernandez E. R., Komatsu E., 2006, ApJ, 646, 703Fernandez E. R., Komatsu E., Iliev I. T., Shapiro P. R., 2010, ApJ, 710, 1089Franceschini A., Rodighiero G., Vaccari M., 2008, A&A, 487, 837Freese K., Spolyar D., Aguirre A., 2008, J. Cosmol. Astropart. Phys., 11, 14Gilmore R. C., Somerville R. S., Primack J. R., Domınguez A., 2011,

preprint (arXiv:1104.0671) (GSPD11)Gorjian V., Wright E. L., Chary R. R., 2000, ApJ, 536, 550Gould R. J., Schreder G. P., 1967, Phys. Rev., 155, 1404

C" 2011 The Author, MNRAS 420, 800–809Monthly Notices of the Royal Astronomical Society C" 2011 RAS

Star Formation Rate Density(Madau Plot)

2σ3σ

5σUpper bounds on the redshift z = 6 - 9 Pop-III SFRD in two possible scenarios with future Fermi GRBs, in the Larson IMF case. The solid lines show the limits from a GRB with the same redshift and spectral characteristics of GRB 080916C (z = 4.35), but with a highest energy observed photon of 30 GeV (160 GeV as emitted) instead of 13.2 GeV, in combination with the 5 most constraining z ≳ 2 sources (Abdo+2010). The dotted lines show a case with a GRB at z = 7 and a highest energy observed photon at 15 GeV (120 GeV emitted).

New data on attenuation of gamma rays from blazers ● X-ray + Fermi + ACT SSC fits to 9 blazars (Dominguez+12) ● Fermi data on 150 blazars at z = 0 - 1.6 (Ackermann+12) now lead to statistically significant measurements of the cosmic gamma ray horizon and EBL as a function of source redshift and gamma ray energyThese new measurements are consistent with recent EBL calculations based both on multiwavelength observations of thousands of galaxies and also on semi-analytic models of the evolving galaxy population. Such comparisons account for all the light, including that from galaxies too faint to see. Catching a few high-redshift GRBs with Fermi or low-threshold atmospheric Cherenkov telescope arrays could provide important new constraints on the high-redshift star formation history of the universe. Happy Birthday Felix!

Conclusions

Documents

The Detection of the Cosmic γ-ray Horizon The Detection of thephysics.ucsc.edu/~joel/Primack-EBL-Barcelona.pdf · Fourth International Fermi Symposium, Monterey, CA October 28 –