Exploring the Universe at the highest energies with the Pierre...

Exploring the Universeat the highest energies with the Pierre Auger ObservatoryKarl-Heinz KampertUniversity Wuppertal

OBSERVATORY

1 Photo by Steven SaffiOdense (Denmark), May 19-22, 2014

Karl-Heinz Kampert - Pierre Auger Observatory Mass 2014, Odense (Denmark), May 19-22, 20142

1 event per m2 and sec

γ≈ 2.7 - 3.0

„ankle“(1 per km2-year)

log(energy/eV)

log(

flux)

32 o

rder

s of

mag

n.

„Knee“(1 per m2-year)

Satellite <1 m2

Balloons <5 m2

0.1 km2

Air Shower Exps

3000 km2


courtesy R. Engel

Particle Energy (eV)

1310 1410 1510 1610 1710 1810 1910 2010

)1.

5 e

V-1

sr

-1 s

ec-2

J(E

) (m

2.5

Scal

ed fl

ux

E

1310

1410

1510

1610

1710

1810

1910

(GeV)ppsEquivalent c.m. energy 210 310 410 510 610

RHIC (p-p)-p)eHERA (

Tevatron (p-p)LHC (p-p)

ATICPROTONRUNJOB

KASCADE (QGSJET 01)KASCADE (SIBYLL 2.1)KASCADE-Grande (prel.)Tibet ASg (SIBYLL 2.1)

HiRes-MIAHiRes IHiRes IIAuger SD 2008

SNR?

Diffusion lossesfrom Galaxis ?

SNR ?

Galactic CRs?

magn. confinement➠ Emax ~ Z

Features of CR spectrum

Fe-knee

p,He-knee

AGN ?

Extragal. CRs?

pFe

? extragal.component


Lamor radii at 1020 eV compared to Milky-Way

E18 ≤ Z·BµG·Rkpc

Conjecture:Extragalactic

origin

Size × B-Field needsto be very large …

Interesting feature:Can do astronomy with cosmic rays !

1020 eV CRs in our Galaxy ?

6


Cosmic Magnetic Fields

Halo B?

Extra-galactic B < nG ?

γ, ν

weak deflection

RL = kpc Z-1 (E / EeV) (B / μG)-1

RL = Mpc Z-1 (E / EeV) (B / nG)-1

strong deflection

Milky wayB ~ μG

E > 1019eV

E < 1018eV

�(E,Z) ⇥ 0.8��

1020 eVE

⇥ ⇤L

10 Mpc

⇤Lcoh

1 Mpc

�B

1 nG

⇥· Z

UHECR Astronomy


Active GalacticNuclei (AGN)

LHC

GRB

AGN-Jets

SNR

Colliding Galaxies

Potential Sources of 1020 eV particles

8

Neutron Stars

white dwarfts

Active Galactic Nuclei ?

jets from radioInterplanetary

Space

Galact.disk

halo

eV proton

galaxies

GalacticClusters

Size

Mag

net

ic F

ield

stre

ng

th (G

auß

)

1AU

SNR

1012

10 6

1

10 –6

1km 10 6 km 1pc 1kpc 1Mpc

IGM

10 20

{

LHC

GRB ?Emax ~ βs·z·B·L

Fe

Hillas Diagramm

Realistic constraints more severe

• small acceleration efficiency• synchrotron & adiabatic losses• interactions in source region


Cas A(3.4 kpc)

Cygnus A(250 Mpc)

Fornax A (20 Mpc)

NRAO/AUI

1.4 , 5, & 8.4 GHz

Supernova RemnantsAccreting

Supermassive Black Holes

E < 1016 eV

E ~ 1020 eV ?

Radio Images of Cosmic Accelerators

9


• Where do they come from?

• What are they made of ?

• How do their accelerators work?

• Is there a maximum limit to their energy ?

• What can can they tell us aboutfundamental and particle physics?

Key Questions aboutUltra High-Energy Cosmic Rays


Hybrid Observation of Extensive Air Showers

11

Particle-density and-composition at ground

light traceat night-sky(calorimetric)

Also:Detection of Radio- & Microwave-Signals

Fluorescence light

Concept pioneered by thePierre Auger Collaboration(Fully operational since 06/2008)


Pierre Auger Observatory

12

3000 km2

~65 km

~65 k

m

CoihuecoHEAT

BLS

CLF

XLF

Loma Amarilla

Los Morados

Los Leones

1660 detector stationson 1.5 km grid

27 fluores. telescopesat periphery

130 radio antennas...Province Mendoza, Argentina


SANTIAGO

A R G E N T I N A

Pampa AmarillaProvince of Mendoza1400 m a.s.l.35° South, 69° West3000 km2


3000 km2 area, Argentina27 fluorescence telescopes plus

...1660 Water Cherenkov tanks

Auger Hybrid Observatory


Pierre Auger Collaboration~500 Collaborators; 90 Institutions, 18 Countries:Argentina Poland UK

Australia Portugal USA

Brasil Romania

Czech Republic Slovenia Bolivia*

France Spain Vietnam*

Germany

Italy USA

Mexico

Netherlands

16

*Associated

OBSERVATORY

☀

Full membersAssociate members

☀


Event Example in Auger Observatory

12 km

~ 20 km

OBSERVATORY


Event Example in Auger Observatory

12 km

~ 20 km

E = 68 EeVXmax=770 g/cm2

dE/

dx

(PeV

/g c

m2 )

Slant Depth (g cm2)

0

40

400 600 800 1000 2000

80

120

160Longitudinal Profile

Lateral Profile

1000

S(1000)

S(1000)=222 VEMθ = 54°S38=343 VEME = 71 EeV

10

100

1000

Sign

al (V

EM)

Distance to Shower Core (m)

12000 3000

OBSERVATORY

Cross Correlation

Infill

Standard

inclined

calorimetric meas.

µ+e measurement

17.5 18.0 18.5 19.0 19.5 20.0 20.5log10( E /eV )

1036

1037

1038

E3J(E)

eV2km

−2sr

−1yr

−1

1035

863

1736

5622

0112

9532

4226

2720

1514

1052

202

2968

421

413

1301

486

2458

0739

8427

0017

0111

1667

642

718

890

457

31

1018 1019 1020E [eV ]

Auger 2013 preliminary


OBSERVATORY

End of the CR-Spectrum

γ1=3.23±0.07 γ2=2.63±0.04

Eankle=5·1018 eV

* *E50% =4·1019 eV

130 000 events

J(E;E > Ea) µ E�g2

1+ exp

✓ log10 E � log10 E1/2

log10 Wc

◆��1

g g

29

Is this the GZK-effect... ?


threshold: EpEγ > (mΔ2 - mp2)

⇒ EGZK ≈ 6·1019 eV100

101

102

103

104

1018 1019 1020 1021 1022

Ener

gy L

oss

Path

leng

th (M

pc)

Energy (eV)

Expansion

CMBRz = 0

Photopion, p

Photopair, p

Total, p

Photopair,Fe

Photopion,Fe

ePhotodisintegration, F

20

Greisen-Zatsepin-Kuz‘min (1966)

GZK-Effect: Energy losses in CMB

p

CMB

pπ

A

CMB

➙ GZK-Horizon ~ 60 Mpc

p+ �CMB ! � ! p+ ⇡0photo-pion production

A+ �CMB ! (A� 1) + n...photo disintegration

17.5 18.0 18.5 19.0 19.5 20.0 20.5log10(E/eV)

1036

1037

1038

E3 J

(E)[e

V2

km-

2sr

-1

yr-

1 ] �E/E = 14 %

Proton, Ecut = 1020 eVProton, Ecut = 1020.5 eVIron, Ecut = 1020 eVIron, Ecut = 1020.5 eV

1018 1019 1020E [eV]


Data compared to GZK-effect

(Phys. Rev. Lett., 2008, updated)

p-sources

Fe-sources

lg (E [eV])18 18.5 19 19.5 20 20.5

]-1

yr

-1 s

r-2

km

2 [e

V3

diff

flux

X E

3710

3810

all

p

He

=3-12obsZ

=12-26obsZ

primary p

Auger ICRC2013


Protons Emax,p = 1018.9 eV

Model inspired by Allard, Astropart. Phys. 39-40, 2012

Limiting Energy of Sources (Emax~Z) + GZK

Simulations done with CRPropa 2.0

m=0; γ=1.55In this case GZK-effect isnot responsible for cut-off!

Iron Emax, Fe = 26 Emax,p = 1020.3 eV

]2slant depth [g/cm200 400 600 800 1000 1200 1400 1600

)]2 e

nerg

y de

posi

t [Pe

V/(g

/cm

0

10

20

30

40

50 Auger event


Longitudinal Shower Development ➙ Primary Mass

23

OBSERVATORY

Example of a 3·1019 eV EAS event in FD

KHK, Unger, APP 35 (2012)EPOS 1.99 Simulations

<Xmax>

RMS(Xmax)


Xmax and RMS(Xmax) as a fct of E

24

Updated Measurement of hXmax

i and �(Xmax

)

comparison to air shower simulations

E [eV]1810 1910 2010

]2 [g

/cm

〉m

axX〈

660

680

700

720

740

760

780

800

820

840

proton

iron

EPOS-LHC QGSJetII-04 Sibyll2.1

E [eV]1810 1910 2010

]2) [

g/cm

max

(Xσ

0

10

20

30

40

50

60

70

80

90 Auger 2013 preliminary

proton

iron

[8 of 15]

OBSERVATORY

OBSERVATORY

Auger data show a smooth changeto a heavier composition above 5 EeV

using post LHC interaction models:

3844

64

96

129

174

246

324

431

lg(E [eV])18 18.5 19 19.5 20 20.

]-1

yea

r-1

sr

-2 k

m2

[eV

3di

ff flu

x * E

3710

3810

Auger ICRC2013


Composition compared with astrophys. scenarios

25

]2>

[g/c

mm

ax<X

650

700

750

800

850

900Proton

Iron

proton source

iron source

Auger ICRC2013

lg(E [eV])18 18.2 18.4 18.6 18.8 19 19.2 19.4 19.6 19.8 20 20.2

]2) [

g/cm

max

(Xσ

0

10

20

30

40

50

60

70Proton

Iron

Auger ICRC2013

proton source

iron source

p

FeGCR sourcecomposition

GCR sourcecomposition

]2>

[g/c

mm

ax<X

650

700

750

800

850

900Proton

Iron

proton source

iron source

Auger ICRC2013

lg(E [eV])18 18.2 18.4 18.6 18.8 19 19.2 19.4 19.6 19.8 20 20.2

]2) [

g/cm

max

(Xσ

0

10

20

30

40

50

60

70Proton

Iron

Auger ICRC2013

proton source

iron source

lg (E [eV])18 18.5 19 19.5 20 20.5

]-1

yr

-1 s

r-2

km

2 [e

V3

diff

flux

X E

3710

3810

all

p

He

=3-12obsZ

=12-26obsZ

primary p

Auger ICRC2013

Limiting energy of sources combined with

GZK describes composition data best


Mass Fractions from Xmax DistributionsSlide removed (data not yet public)


Implications of a heavy composition

27

Extragalactic propagation of ultrahigh energy cosmic-rays q

Denis Allard

Laboratoire Astroparticule et Cosmologie (APC), Université Paris 7/CNRS, 10 rue A. Domon et L. Duquet, 75205 Paris Cedex 13, France

a r t i c l e i n f o

Article history:

Available online 10 November 2011

a b s t r a c t

In this paper we review the extragalactic propagation of ultrahigh energy cosmic-ray (UHECR). We pres-

ent the different energy loss processes of protons and nuclei, and their expected influence on energy evo-

lution of the UHECR spectrum and composition. We discuss the possible implications of the recent

composition analyses provided by the Pierre Auger Observatory. The production of secondary cosmogenic

neutrinos and photons and the constraints their observation would imply for the UHECRs origin are also

addressed. Finally, we conclude by briefly discussing the relevance of a multi messenger approach for

solving the mystery of UHECRs. ! 2011 Elsevier B.V. All rights reserved.

1. Introduction

After more than 50 yr of experimental efforts, the origin of ultra-

high energy cosmic-rays (UHECRs, e.g., cosmic-rays above

!1018 eV) remains a mystery. The understanding of the production

of these particles, the most energetic particles in the universe, is

one of the most intense research field of high energy astrophysics.

Since the pioneering experiment at Volcano Ranch and the observa-

tion of the first cosmic-ray event above 1020 eV (see [1] for a com-

plete review of the early experiments), large statistics have been

accumulated above 1018 eV. High resolution measurements of the

UHECR spectrum, composition and arrival direction have been al-

lowed by recent experiments like AGASA [2], HiRes [3], the Pierre

Auger Observatory [4] and Telescope Array [5]. Among the most

interesting recent results, one can cite (see [6]) the evidence for a

suppression of the UHECR flux above 3–5 " 1019 eV observed by

HiRes [7] and Auger [8] with a large significance. Furthermore,

the recent analyses at the Pierre Auger Observatory seem to indi-

cate an evolution of the composition toward heavier elements

above the ankle [9] as well as hints for an anisotropic distribution

of the arrival directions of the highest energy events [10] and in par-

ticular a possible diffuse excess in the direction of the Centaurus

constellation[11]. Since the statistics above !3 " 1019 eV are quite

low, and the consistency between the results of different experi-

ments is still a matter of debate, these trends remain to be con-

firmed and understood with future data.

The extragalactic origin of UHECRs (at least above the ankle of

the cosmic-ray spectrum) is widely accepted. As a consequence

the measured cosmic-ray spectrum on Earth has to be shaped by

the effect of the propagation of the particles in the extragalactic

medium. During their journey from the source to the Earth the in-

jected cosmic-ray spectrum and composition can be modified by

interactions with photon backgrounds and cosmic magnetic fields.

A detailed modeling of the extragalactic propagation of UHECRs is

then a necessary ingredient for the astrophysical interpretation of

the data. One of the most important features due to UHECRs extra-

galactic propagation is the prediction of a cut-off in the observed

spectrum above a few 1019 eV due to interactions of UHE protons

or nuclei with photons of the Cosmic Microwave Background

(CMB). This prediction [12,13] of the so-called GZK cut-off (named

after the authors of the original studies Greisen, Zatsepin and Kuz-

min) was made in 1966, closely following the discovery of the

CMB. This prediction started a long series of studies on the extra-

galactic propagation of UHECRs including the production of sec-

ondary neutrinos and photons.

In this paper, we review the main aspects of the extragalactic

propagation of UHECR protons and nuclei. It will be organized as

follows. In the next section, we will review the main interaction

channels of protons and nuclei and discuss their influence on the

energy and mass losses. In Section 3., we show some propagated

spectra, allowing us to discuss expectations concerning the evolu-

tion of the composition from the source to the Earth and the pro-

duction of secondary protons. We discuss, in particular, the

possible implications of the composition trend suggested by the

recent analyses of the Pierre Auger Observatory. In Section 4 we

discuss discuss the production of secondary messengers (neutrinos

and photons) and the possible constraints their observation could

bring for the understanding of the UHECR origin. in Section 5.,

we finally conclude by briefly discussing the prospects for improv-

ing our understanding of the UHECR phenomenon and the ex-

pected contributions of current and planed experiments in

cosmic-rays, gamma-rays and neutrinos.

0927-6505/$ - see front matter ! 2011 Elsevier B.V. All rights reserved.

doi:10.1016/j.astropartphys.2011.10.011

q (A section discussing the influence of extragalactic magnetic fields and the

possible 14 departures from the rectilinear case is available in an extended version

of the present review available from: <arXiv:1111.3290>.)

E-mail address: [email protected]

Astroparticle Physics 39–40 (2012) 33–43

Contents lists available at SciVerse ScienceDirect

Astroparticle Physics

journal homepage: www.elsevier .com/ locate/ast ropart

UHECR composition models

Andrew M. Taylor⇑

Dublin Institute for Advanced Studies, 31 Fitzwilliam Place, Dublin 2, Ireland


Article history:

Received 11 July 2013

Received in revised form 4 November 2013

Accepted 8 November 2013

Available online 16 November 2013

Keywords:

UHECR

Composition

a b s t r a c t

In light of the increasingly heavy UHECR composition at the highest energies, as observed by the Pierre

Auger Observatory, the implications of these results on the actual source composition and spectra are

investigated. Depending on the maximum energy of the particles accelerated, sources producing hard

spectra and/or containing a considerably enhanced heavy component appear a necessary requirement.

Consideration is made of two archetypal models compatible with these results. The secondary signatures

expected, following the propagation of the nuclear species from source to Earth, are determined for these

two example cases. Finally, the effect introduced by the presence of nG extragalactic magnetic fields in

collaboration with a large (80 Mpc) distance to the nearest source is discussed.

! 2013 Elsevier B.V. All rights reserved.

1. Introduction

During the last decade the field of UHECR research has under-

gone considerable developments with the completion of extremely

large detector facilities. The data from these instruments has lead

to a notable improvement in both the quantity and quality of

UHECR measurements. Following the digestion of this new infor-

mation, a revision of the UHECR model working hypothesis may

be due. In particular, measurements sensitive to the UHECR com-

position have improved dramatically with a coherent picture start-

ing to emerge from the ensemble of different composition sensitive

measurements the Pierre Auger Observatory (PAO) has made [1]. It

should be noted that this picture is obscured somewhat when

additional observational data from the TA experiment are included.

The statistical significance of this disagreement, however, remains

unclear. In this study, such additional observational data sets are

neglected.

2. Monte Carlo modeling

In order to test different hypothesis models, a Monte Carlo

description of UHECR propagation is used, as first described in

[2]. In this description, UHECR protons and nuclei are propagated

through the cosmic microwave background (CMB) and cosmic

infrared background (CIB) radiation fields, undergoing photo-disin-

tegration, photo-pion, pair production, and redshift losses as they

do so. Though the cross-sections and target photon spectral distri-

butions relevant for the proton related energy loss processes are

well understood, some uncertainty still remains in both the

photo-disintegration cross-sections and the CIB spectral distribu-

tion relevant for nuclei propagation. In the present study, the

description of these adopted are [3,4] for the cross-sections and

CIB spectral distribution respectively. In Sections 3–5, extragalactic

magnetic field (EGMF) effects are neglected. The effects introduced

by such fields on the main results are discussed in Section 6. In or-

der to take account of EGMF effects, the ‘‘delta-approximation’’

prescription provided in [5] is implemented.

To perform a comparison with the PAO measurements, the

predicted values of the composition sensitive shower profile

parameters, Xmax and RMSðXmaxÞ, were determined for each model.

In order to encapsulate the uncertainty in the hadronic model

description for these values, the spread in predicted values from

four different models [6–9] was determined.

The Monte Carlo description was applied to an ensemble of dis-

tributed sources whose redshift evolution scaled as ð1þ zÞm , with

m ¼ 3 from zmin (with corresponding nearest source distance

Lmin) up to zmax ¼ 1:5. An energy spectrum output by each source,

of the form dN=dE / E%ae%ðE=Emax;ZÞ , with Emax;Z ¼ ðZ=26ÞEmax;Fe, was

adopted.1 Source spectral indices in the range 1 < a < 3, and 3-com-

ponent compositions were scanned over for both the cutoff energy

cases of Emax;Fe ¼ 1020 eV and Emax;Fe ¼ 1020:5 eV. Only spectral and

composition data points with energies above 1018:6 eV were used

in the analysis. The systematic errors for the energy resolution,

Xmax, and RMSðXmaxÞ, were also included in the v2 determination.

The regions of parameter space for which good fits to both the

spectral and composition data were found are shown in Fig. 1.

From each of the two cutoff energy results, two example models

0927-6505/$ - see front matter ! 2013 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.astropartphys.2013.11.006

⇑ Tel.: +353 16621333x337.

E-mail address: [email protected]

1 Such a parameterisation should be considered to provide ‘‘effective parameters’’,

which broadly encapsulate the true emission spectrum properties.

Astroparticle Physics 54 (2014) 48–53

Contents lists available at ScienceDirect



Needfor a

localsourc

e of ultrahi

gh-energy c

osmic-raynucle

i

Andrew M. Tayl

or,1 Markus

Ahlers,2 and Felix

A. Aharonia

n3,4

1 ISDC, Che

min d’Ecogia 16, Ve

rsoix,CH-1

290, SWITZER

LAND

2C. N.Yang

Institute for

Theoretical

Physics, SU

NY at Stony Br

ook, Stony B

rook,New York

11794-3840

, USA

3Dublin Institu

te forAdvan

ced Studies, 5 Merrion

Square, Du

blin 2, IRELAND

4Max-Planck-

Institut fur K

ernphysik,

Postfach 10398

0, D-69029

Heidelberg,

GERMANY

(Received 11 July 2011;

published 7 Nove

mber 2011)

Recent resu

lts of the PierreAuge

r (Auger) fluore

scence detec

tors indicate an increa

singlyheavy

composition of ult

ra-high energ

y (UHE) co

smic rays (CR

s). Assuming th

at thistrend

continues u

p to the

highest ene

rgiesobser

ved by the

Auger surf

ace detector

s, wederive

the constrai

nts this pla

ces onthe lo

cal

source distrib

utionof UH

E CR nuclei. Uti

lizingan analy

tic description

of UHE CR propa

gation, we

derivethe expec

ted spectra and compositi

on for awide

rangeof sou

rce emissionspectr

a. We find that

sources of

intermediate

-to-heavy nucle

i areconsi

stentwith

the observed spectr

a and composition data

abovethe an

kle. This co

nsistency re

quiresthe pr

esence of n

earbysourc

es within 60

Mpc and 80 M

pc for

silicon and iron-o

nly sources, re

spectively.

The necessity of the

se localsourc

es becomes even

more

compelling in the pr

esence nan

o-Gauss loc

al extragala

ctic magnet

ic fields.

DOI:10.11

03/PhysRev

D.84.10500

7

PACSnumbers:

13.85.Tp, 9

8.70.Sa

I. INTRODUC

TION

Ultra-high

energy CRs

withenerg

ies above10

19 eV

arriveat Ea

rth witha frequ

encyof les

s thanone event

per square-kilo

meter-year (i.e.

withan energ

y fluxof

30 eV cm!2 s!

1 ) in ! steradian. T

he recently completed

PierreAuge

r observator

y [1]with a

surface dete

ctor cover-

ing an area of abou

t 3000km

2 is thus able

to detect up to

several hund

red UHE CR event

s peryear.

However, d

ue to

the steeply

falling CR

spectrum the ar

rival freque

ncy drops

by about2 order

s of magnitu

de aswe go

up in energy by

one decade

, leaving only a few

events per

year detecta

ble

at energies

around 10

20 eV. Hence,

the statistical u

ncer-

taintyassoc

iatedwith t

he upper en

d of the spec

trumis stil

l

large,limiting

our knowledge

of UHECR compositi

on

and origin.

Before thei

r arrival, UH

E CRs must pr

opagate acr

oss the

astronomical d

istance betw

een their so

urce and Ea

rth. The

relevant inte

ractions of

UHECR nucle

i during pro

pagation

are Bethe-H

eitlerpair p

roduction a

nd photo-d

isintegratio

n

in collision

s withphoto

ns ofthe co

smic backgrou

nd radia-

tion. The cro

ss sections

of both thes

e processes

rise quickly

abovethe thresh

old values of a

bout1 MeV and 10 MeV,

respectivel

y, inthe nucle

i’s rest-frame. At

evenhighe

r

energies above

150 MeV pionprodu

ctionturns

on and

becomes the

dominantenerg

y loss proces

s. However

, for

the UHE CR cutoff

energies an

d compositi

on weconsi

der,

whichare m

otivated by

the most re

cent Auger r

esults, this

process never

playsa dominant

role and may be safely

neglected.

Provided the propa

gationtime from

theirsourc

es to

Earthis greate

r thantheir

energy loss time, UH

E CRs

invariably

undergo these

energy loss

interactions

. The

breakup of nu

clei via pho

to-disintegr

ationprodu

ces lower

mass nuclei with

the same Lorentz factor

. For heavy

nuclei, the most dominant

transitions

are one-nucleo

n

and two-nucleo

n losses. Sec

ondary heavy

nuclei rem

ain

closeto the photo

-disintegrat

ion resonance

and quickly

disintegrate

further to li

ghternucle

i. Hence, on

resonance,

the initialmass compositi

on of the sources is quick

ly

shifted to lower

atomic number values and in gener

al

showsa strong

dependence

on CR energy.

The arrivin

g fluxfrom

an ensemble of

UHECR sourc

es

can be exp

ectedto con

tain suppre

ssionfeatur

es at the hig

h

energies at

whichthe photo

-disintegrat

ion processes t

urn

on and the nuclei part

icle’sattenu

ationlength

decreases.

Most prominentl

y, for the caseof a proto

n-dominated

spectrum, the

flux is expected to be suppr

essedby the

Greisen-Za

tsepin-Kuz’

min (GZK)

cutoff[2,3]

due toreso-

nantpion

photo-prod

uctionintera

ctionswith

the cosmic

microwave ba

ckground (C

MB). Intrigui

ngly,a supp

ression

of theCR spectr

um at about 5

" 1019 eV has been

ob-

servedat a statist

icallysignifi

cant level [

4,5].Howe

ver,

sincea similar fe

aturemay also appea

r in the spectrum

fromnucle

i primaries,

the observation

of such a featur

e

provides litt

le clue as to

the underlyi

ng compositi

on. Such a

suppression

feature becom

es a cutoffin the arrivi

ng flux

if theattenu

ationlength

dropsbelow

the distance to the

nearest UH

ECR source. Thu

s, theshape

of thesuppr

ession/

cutofffeatur

e doesconta

in information

aboutthe sourc

e

distribution,

as hasbeen i

nvestigated

already for t

he case of

UHECR proto

ns [6–8].

On theirarriva

l at Earth,

UHECRs i

nteract with

mole-

culesin the atmosphe

re and deposit the

ir energy in the

formof ext

ensiveair sh

owers. The

characteris

tics of thes

e

showers alo

ng theshowe

r depth X (in g=

cm2 ) cont

ain vital

UHECR compositi

on information.

On average, p

roton-

induced showe

rs reachtheir

maximum development,

hXmaxi, deeper i

n the atmosphere than do showe

rs ofthe

same energy gener

ated by heavier nu

clei. Accom

panying

this effect,

the shower

to shower fl

uctuation of Xmax

about

PHYSICAL

REVIEW D 84, 10

5007(2011

)

1550-7998=

2011=84(10

)=105007(1

0)

105007-1

! 2011American

Physical So

ciety

On the heavy chemical composition of the ultra-high energy cosmic rays

Dan Hooper a,b, Andrew M. Taylor c,d,*

a Fermi National Accelerator Laboratory, Theoretical Astrophysics, Batavia, IL 60510, USA

b University of Chicago, Department of Astronomy and Astrophysics, Chicago, IL 60637, USA

c Max-Planck-Institut für Kernphysik, Postfach 103980, D-69029 Heidelberg, Germany

d ISDC, Chemin d’Ecogia 16, Versoix, CH-1290, Switzerland


Article history:

Received 11 October 2009

Received in revised form 10 December 2009

Accepted 10 January 2010

Available online 18 January 2010

Keywords:

UHECRComposition

a b s t r a c t

The Pierre Auger Observatory’s (PAO) shower profile measurements can be used to constrain the chem-

ical composition of the ultra-high energy cosmic ray (UHECR) spectrum. In particular, the PAO’s measure-

ments of the average depth of shower maximum and the fluctuations of the depth of shower maximum

indicate that the cosmic ray spectrum is dominated by a fairly narrow distribution (in charge) of heavy or

intermediate mass nuclei at the highest measured energies ðE J 1019 eVÞ, and contains mostly lighter

nuclei or protons at lower energies ðE # 1018 eVÞ. In this article, we study the propagation of UHECR

nuclei with the goal of using these measurements, along with those of the shape of the spectrum, to con-

strain the chemical composition of the particles accelerated by the sources of the UHECRs. We find that

with modest intergalactic magnetic fields, 0.3 nG in strength with 1 Mpc coherent lengths, good fits to the

combined PAO data can be found for the case in which the sources accelerate primarily intermediate

mass nuclei (such as nitrogen or silicon). Without intergalactic magnetic fields, we do not find any com-

position scenarios that can accommodate the PAO data. For a spectrum dominated by heavy or interme-

diate mass nuclei, the Galactic (and intergalactic) magnetic fields are expected to erase any significant

angular correlation between the sources and arrival directions of UHECRs.! 2010 Published by Elsevier B.V.

1. Introduction

The chemical composition of the ultra-high energy cosmic ray

(UHECR) spectrum has long been a topic of great interest [1–6].

Until recently, however, very little was known about the nature

of these particles. On one side of the debate, the so called Hillas cri-

terion [7] gives a preference for the electromagnetic fields of cos-

mic ray sources to accelerate heavy nuclei to higher energies

than protons or light nuclei. On the other side, it has been argued

that the angular correlations reported by the Pierre Auger Observa-

tory (PAO) [8], as well as features in the shape of the UHECR spec-

trum [9], suggest that these particles consist largely of protons.

None of these arguments, however, has yet settled the question

of what types of particles make up the UHECR spectrum.

Data from the Pierre Auger Observatory (PAO), however, is

offering increasingly powerful insights into this question. Firstly,

the spectral shape predicted for the UHECR all-particle spectrum

depends not only on the injected spectrum and spatial distribution

of the sources, but also on the chemical composition that is in-

jected from the sources of the highest energy cosmic rays. As the

PAO measures the UHECR spectrum with increasing precision

[10], this information can be used to constrain the chemical com-

position of these particles [11]. Furthermore, the PAO is capable

of performing several measurements that can be used to directly

or indirectly determine the chemical composition of UHECRs as

they enter the Earth’s atmosphere. Among these empirical tools

are the measurements of the average depth of shower maximum,

hXmaxi, and the RMS variation of this quantity, RMSðXmaxÞ. On aver-

age, proton-induced showers reach their maximum development,

hXmaxi, deeper in the Earth’s atmosphere than do showers of the

same energy generated by heavier nuclei. Accompanying this re-

sult, the shower to shower fluctuation of Xmax about the mean,

RMSðXmaxÞ, is larger for proton-induced showers than for iron-in-

duced showers of the same energy. As a result, measurements of

both hXmaxi and RMSðXmaxÞ can be used to infer the average chem-

ical composition of the UHECRs as a function energy.

Very recently, the PAO collaboration has announced their first

measurements of RMSðXmaxÞ [12,13]. These measurements, along

with those of hXmaxi, imply that the UHECR spectrum contains a

large fraction of heavy or intermediate mass nuclei, especially at

the highest energies measured. Furthermore, the small values of

RMSðXmaxÞ measured by the PAO also imply that the composition

of the UHECR spectrum is relatively narrowly distributed at the

highest measured energies, containing little or no protons or light

0927-6505/$ - see front matter ! 2010 Published by Elsevier B.V.

doi:10.1016/j.astropartphys.2010.01.003

* Corresponding author. Address: Max-Planck-Institut für Kernphysik, Postfach

103980, D-69029 Heidelberg, Germany.

E-mail address: [email protected] (A.M. Taylor).


Contents lists available at ScienceDirect



arX

iv:1

312.

7459

v1 [

astro

-ph.

HE]

28

Dec

201

3

Prepared for submission to JCAP

Ultra high energy cosmic rays:implications of Auger data for sourcespectra and chemical composition

R. Aloisio1,2, V. Berezinsky2,3 and P. Blasi1,2

1INAF/Osservatorio Astrofisico di Arcetri, Largo E. Fermi, 5 - 50125 Firenze, Italy2Gran Sasso Science Institute (INFN), viale F. Crispi 7, 67100 L’Aquila, Italy3INFN/Laboratori Nazionali Gran Sasso, ss 17bis km 18+910, 67100 Assergi, Italy

E-mail: [email protected],[email protected], [email protected]

Abstract. We use a kinetic-equation approach to propagation of ultra high energy cosmicray protons and nuclei to infer possible implications of the data on spectrum and chemicalcomposition collected by the Pierre Auger Observatory. Using a homogeneous source dis-tribution, we show that a simultaneous fit to the spectrum, elongation rate Xmax(E) anddispersion σ(Xmax) implies the injection of nuclei with very hard spectra. This leads howeverto underestimate the flux at energies E ≤ 5× 1018 eV, thereby implying that an additionalcosmic ray component is required, which needs to be of extragalactic origin. We discuss thenature of this additional component in terms of the recent findings of KASCADE-Grandeon fluxes and chemical composition, which allows to describe the transition from Galactic toextragalactic cosmic rays.

Astroparticle Physics 54, 48 (2014)

Astroparticle Physics 39–40 (2012) 33–43


Subm. to JCAP 2013

...and many more papers of this type

all require very hard injection spectra unless

a nearby source (population) is assumed

E [eV]1810 1910 2010

]-1

y-1

sr

-2in

tegr

al fl

ux [k

m

-510

-410

-310

-210

-110

1

10 SHDM GZK pTD GZK FeZ-burst

AGASAYakutskAuger SDAuger HybridTA SD

Cosmogenic+neutrinos+

4+

Detec?on+in+EeV+range+may+provide+complementary+informa?on+to+direct+UHECR+detec?on+on:++UHECR+nature+(p,+mixed,+Fe),+origin+(evolu?on+of+the+sources,+maximum+energy+a^ainable,…)+

[eV]νE1710 1810 1910 2010 2110

]-1

sr

-1 s

-2 d

N/d

E [

GeV

cm

2E

-1010

-910

-810

-710

-610

-510

-410Single flavour

p, best fit with Fermi-LAT bound (Ahlers)

eV (B. Sarkar)20=10maxp, FRII & SFR, E

eV (B. Sarkar)20=26 x 10maxFe, FRII & SFR, E

eV (Kotera)21.5 - 1020=Z x 10maxp & mixed, SFR & GRB, E

p+ �CMB ! � ! p+ ⇡0


Smoking Gun of GZK-effect

28

! p+ ⇡0! �EeV! n+ ⇡+! ⌫EeV

cosmogenic neutrinos and photons– a guaranteed signal –

GZK-p

GZK-Fe

GZK-p

GZK-Fe

TopDown models

Energy[eV]1810 1910 2010

]-1

y-1

sr

-2 [k

m0

Inte

gral

Flu

x E>

E

-310

-210

-110

1upper limits 95% C.L.

SD

Hyb 2011

Hyb 2009A

A

YY

TA

SHDMSHDM'TDZ-burstGZK


Diffuse Photon Limits

GZK

Photon upper limits rule out Top-Down Modelsand get close to expected GZK-fluxes

OBSERVATORY

Sensi?vity+to+all+flavours+&+channels+

��

��

��

��

��

��

��

��

�

��

��

�

��

� ��

��

��!��!� ��"� �#�$��!�� "%�$� �#��

10+


Sensitivity to all ν flavors and channels


Differential diffuse ν flux limits

31

[eV]E1710 1810 1910 2010 2110

]-1

sr

-1 s

-2 d

N/d

E [

GeV

cm

2 E

-1010

-910

-810

-710

-610

-510

-410Single flavour (90% CL)

Pierre AugerPRELIMINARY

[eV]E1710 1810 1910 2010 2110

]-1

sr

-1 s

-2 d

N/d

E [

GeV

cm

2 E

-1010

-910

-810

-710

-610

-510

-410 limits

Full IceCube (2 yr)Auger (6 yr)ANITA-II (28.5 days)

modelsCosmogenic p, Fermi-LAT boundFe, FRII & SFR evol.p & mixed

Astrophysical sourcesAGN Waxman-Bahcall

All limits converted to single flavour and given per half a decade of energy

GZK


Taking Profit fromthe extreme energies... (1)


Tests of Lorentz Invariance Violation

33

Galaverni & SiglPRL 100, 021102 (2008) and many others

cascading of UHEphotons suppressed

LIV ➙ may modify photon dispersion relation

⇥2 = k2 + m2 + �nk2(k/MPl)n

➙ affect the threshold for e+e– pair production

➙↳ �� e+e�

expect significant photonfraction above ~ 1019 eV

γ-fractionif LIV

no LIV

p + �CMB � �� p + ⇥0

�1 ⇥ 2.4� 10�15

7 orders of magnitudesbetter than previous limits!CPT tests possible as well

�2 ⇤ �2.4⇥ 10�7

Auger upper limits

Auger:Astropart. Phys. 29 (2008) 234


Bounds on LIV and Smoothness of Class. Space-Time

34

Klinkhamer/Risse; PRD77 (2008) 016002; 117901; Klinkhamer, AIP Conf.Proc.977:181-201,2008

• Conservative limit on any small-scale structure of space:LEP/LHC: ℓ≤10-19 m ≈ ℏc/(1 TeV).

• Use published 27 Auger events + 1 AGASA + 1 Fly‘s-Eye

↪ single scale classical space-time foam at

UHECRs: b ≤ 10-26 m ≈ ℏc/(2·1010 GeV)

• Conjecture: fundamental length scale of quantum space time may be different from Planck length and may be linked to cosmological constant

Observation of 1020 eV eventsproofs absence of Vacuum Cherenkov-Radiation➙ Provides limits on smoothness of space & LIV-effects

by far best (3 to 8 orders of magn.!) existing bounds of Standard Model Extension parameters of nonbirefringent modified Maxwell theory

Results complemented by TeV γ-rays


Taking Profit fromthe extreme energies... (2)

]2 [g/cmmaxX500 600 700 800 900 1000 1100 1200

/g]

2 [

cmm

axdN

/dX

-110

1

10

2 g/cm+2.4-2.3 = 55.8 fR

dPdX1

= 1�int

e�X1/�int

�p�Air = hmAiri�int


p-Air Cross-Section from Xmax distribution

Data: 1018 eV < E < 1018.5 eV

In practice: σp-Air by tuning models to describe Λ seen in data

X1: point of 1st interactionΔX1

ΔXmax ≈ X1

Difficulties:• mass composition can alter Λ• fluctuations in Xmax

• experimental resolution ~ 20 g/cm2

Λint

top of atmosphere

Inelastic Proton-Proton Cross-Section

Standard Glauber conversion + propagation of modeling uncertainties

(Pro

ton-

Prot

on)

[m

b]in

elσ

30

40

50

60

70

80

90

100

110

[GeV]s310 410 510

ATLAS 2011CMS 2011ALICE 2011TOTEM 2011UA5CDF/E710Auger ICRC 2011

QGSJet01QGSJetII.3Sibyll2.1Epos1.99Pythia 6.115Phojet

σinelpp =

!

90 ± 7stat (+9−11)sys ± 1.5Glauber

"

mb√spp = [57 ± 6] TeV

The 1.5mb do not reflect the total theoretical uncertainty, since there are other

models available for the conversion.Ralf Ulrich ([email protected]) Measurement of the Proton-Air Cross-Section with the Pierre Auger Observatory 12 / 19


p-Air and pp Cross section @ √s=57 TeV

37

(Energy/eV)10

log11 12 13 14 15 16 17 18 19 20

Cro

ss s

ectio

n (p

roto

n-ai

r)

[mb]

200

300

400

500

600

700

QGSJet01cQGSJetII.3Sibyll 2.1Epos 1.99

Energy [eV]1110 1210 1310 1410 1510 1610 1710 1810 1910 2010

[TeV]ppsEquivalent c.m. energy -110 1 10 210

Nam et al. 1975Siohan et al. 1978Baltrusaitis et al. 1984Mielke et al. 1994Honda et al. 1999Knurenko et al. 1999ICRC07 HiResAglietta et al. 2009Aielli et al. 2009This work

0.9TeV 2.36TeV 7TeV 14TeV

LHC

Conversion from p-airto p-p cross section by Glauber-approach

σp-Air= (505 ± 22stat (+26)sys ) mb –34

σpp = [92 ± 7stat (+9 )sys ± 7.0Glauber] mb –11

LHC OBSERVATORY

OBSERVATORY

inel

prod

σpp = [133 ± 13stat (+17 )sys ± 16Glauber] mb –20

tot

Auger

Auger Collabora,on, PRL 109, 062002 (2012)

0

200

400

600

800

1000

1200

1400

1e-07 1e-06 1e-05 0.0001 0.001 0.01 0.1 1 10 100 1000

σp-

Air [

mb]

Energy [EeV]

σp-Air vs. Energy

QII-04EPOS-LHCTo Fit RMS

AugerQII-04 (π-Air)

EPOS-LHC (π-Air)


Disentangling Int.-Models from Composition

38

17.5 18.0 18.5 19.0 19.5 20.0 20.5log10(E/eV)

600

650

700

750

800

850

900

hXm

axi[

gcm

-2 ]

Post-LHC modelsProtonIron

17.5 18.0 18.5 19.0 19.5 20.0 20.5log10(E/eV)

0

20

40

60

80

100

�( X

max

)[g

cm-

2 ]

Post-LHC modelsProtonIron

If the composition would be protons at all energies, Xmax and RMS(Xmax) could only be described by a rapidly rising cross section above LHC energies

G. Farrar, priv. comm.


⟨ρµ⟩ and RMS(ρµ) in a mixed composition changes distinctively different with Xmax

as compared to models of a pure p-composition

Disentangling Int.-Models from Composition

0.5

1

1.5

2

2.5

3

600 700 800 900 1000 1100 1200

l µ [m

-2]

Xmax [g/cm2]

EPOS-LHC pEPOS-LHC He

EPOS-LHC N

Models of new physics33RD INTERNATIONAL COSMIC RAY CONFERENCE, RIO DE JANEIRO 2013

2.4

2.6

2.8

3

3.2

650 700 750 800 850 900

ρ µ(1

000)

[m-2

]

Xmax [g/cm2]

CSRPDSPPS

PPS-HE

Figure 5: The density of muons at 1000 m as a functionof Xmax for the four template models. The error bars showthe variance in samples of 800 hybrid events, which is anumber achievable at the PAO and TA.

or decay and, thus, have different correlation strengths be-tween Nµ and Xmax. Fig. 5 shows the average Nµ as a func-tion of Xmax and its variance in ensembles of 800 simu-lated events, for 10 EeV showers at a zenith angle of 38o,for each of the four schematic models. These simulationswere done at a fixed zenith angle and energy, but the samestatistical power could be achieved by normalizing show-ers to a fiducial energy and angle. The negative correla-tion between Nµ and Xmax is strongest when the modifica-tion is made at all energies, and weaker when the modifica-tion is applied only at high energy. The PPS and PPS-HEmodels show a negative correlation while the CSR modelshows almost no correlation. Finally, the PDS model actu-ally shows a positive correlation: showers with a shallowXmax produce fewer muons than showers with a deep Xmax.Thus the models can be discriminated at high significancewith presently realizable datasets.The behavior of the PDS and PPS models demonstrates

that the correlation is sensitive to the energy threshold ofthe pion modifications. This is made explicit in Fig. 6,which compares the average Nµ and Xmax of iron, carbon,and proton initiated showers, for different degrees of pionproduction suppression at high energy. As pion productionsuppression is increased, the relative difference betweenthe mean Nµ in iron and proton showers decreases.

5 ConclusionWe argue that the muon deficit in simulations of air show-ers indicates that the hadronic models are incorrectly pre-dicting the fraction of energy which is transfered to theelectromagnetic sub shower. Changing the π0 energy frac-tion or suppressing pion decay are the only modificationswhich can be used to increase the number of muons atground without coming into conflict with the Xmax obser-vations.We have developed four schematic models of hadronic

interactions, all of which are capable of correctly describ-ing the Xmax distributions and number of muons at ground.They utilize both pure proton and mixed primary composi-tions, which demonstrates the need for models to correctlydescribe all air shower observables before they are used tointerpret the primary mass composition.The fourmodels can be distinguished by observations of

the correlation between the number of muonsNµ and Xmax,

0.9

1

1.1

1.2

1.3

1.4

1.5

660 680 700 720 740 760 780 800 820

ρ µ / ρ µ

(p)

Xmax [g/cm2]

p, EPOS 1.99C, EPOS 1.99

Fe, EPOS 1.9920% π -> B,K45% π -> B,K

Figure 6: The dependence of the density of muons at 1000m, relative to protons, on pion production at high energy.

for an ensemble of hybrid events. This correlation providesa crucial new observable for determining the nature ofUHE air showers. Existing hybrid datasets may already belarge enough to rule out some explanations of the muonexcess.Acknowledgements: The authors are members of thePierre Auger Collaboration and acknowledge with grat-itude innumerable valuable discussions with colleagues.This research was supported by NSF-PHY-1212538.

References[1] T. Abu-Zayyad et al., HiRes-MIA Collab., Astrophys.J. 557 (2001) 686.

[2] J. Allen, for PAO, Proc. 32nd ICRC, Beijing, China 2(2011) 83, arxiv:1107.4804

[3] G. Rodriguez, for PAO, Proc. 32nd ICRC, Beijing,China, 2 (2011) 95, arxiv:1107.4809

[4] B. Kegl, for PAO, ICRC 2013 arxiv:1307.5059.[5] I. Valino, for PAO, ICRC 2013 arxiv:1307.5059.[6] The Pierre Auger Collaboration, Phys. Rev. Lett. 104(2010) 091101.

[7] The Pierre Auger Collaboration, JCAP 02 (2013) 026.[8] The TA Collaboration, Proc. 32nd ICRC, Beijing,China, 2 (246) .

[9] The HiRES Collaboration, Phys. Rev. Lett. 104(2010) 161101.

[10] J. Matthews, Astroparticle Physics 22 (2005) 387.[11] R. Ulrich et al, Phys. Rev. D 83 (2011) 054026.[12] T. Pierog and K. Werner, Phys. Rev. Lett. 101 (2008)171101.

[13] H.-J. Drescher and G. Farrar, Phys. Rev. D 67 (2003)116001.

[14] K. Werner et al, Phys. Rev. C 74 (2006) 044902.[15] G. Farrar and J. Allen, Proc. UHECR 2012, CERN;arXiv:1307.2322

[16] G. Farrar and G. Veneziano, in prep.[17] S. Ostapchenko, Nucl. Phys. B - Proc. Supp. 151(2006) 143.

[18] E.-J. Ahn et al., Phys. Rev. D 80 (2009) 094003.[19] J. Alvarez-Muniz et al., arxiv:1209.6474.[20] R. Engel, D. Heck, and T. Pierog, Annu. Rev. Nucl.Part. Sci. 61 (2011) 467.

[21] P. Younk and M. Risse, Astroparticle Phys, 35(2012) 807.

Allen, Farrar ICRC2013, arXiv:1307.7131

CSR: Chiral Symmetry RestorationPDS: Pion Decay SuppressionPPS: Pion Production Suppression

10 EeV, θ=38°


Same measured Eventwith predicted signals for p and Fe

Interaction Models underestimate Muon-numbers

40

Models underestimate µ-content of EAS by 30-60%

models

independent tests e.g. using showeruniversality yield the same results

Measured event with matching p and Fe-simulations

G.R. Farrar et al., Muon content of hybrid PAO CRs33RD INTERNATIONAL COSMIC RAY CONFERENCE, RIO DE JANEIRO 2013

0

10

20

30

40

50

1 1.2 1.4 1.6 1.8 2

S [V

EM]

sec(θ)

TotalPure Muon

Pure EMEM from µ DecayEM from Had. Jetµ from Photprod.

Figure 4: The contributions of different components to theaverage signal as a function of zenith angle, for stations at 1km from the shower core, in simulated 10 EeV proton airshowers illustrated for QGSJET-II-04. The signal size ismeasured in units of vertical equivalent muons (VEM), thecalibrated unit of SD signal size [18].

where a is the energy scaling of the muonic signal; it has thevalue 0.89 in both the EPOS and QGSJET-II simulations,independent of composition [19].

Finally, the variance of S(1000) with respect to Sresc mustbe estimated for each event. Contributions to the varianceare of two types: the intrinsic shower-to-shower variance inthe ground signal for a given LP, sshwr, and the variance dueto limitations in reconstructing and simulating the shower,srec and ssim. The total variance for event i and primarytype j, is s

2i, j = s

2rec,i +s

2sim,i, j +s

2shwr,i, j.

sshwr is the variance in the ground signals of showerswith matching LPs. This arises due to shower-to-showerfluctuations in the shower development which result invarying amounts of energy being transferred to the EM andhadronic shower components, even for showers with fixedXmax and energy. sshwr is irreducible, as it is independentfrom the detector resolution and statistics of the simulatedshowers. It is determined by calculating the variance in theground signals of the simulated events from their respectivemeans, for each primary type and HEG; it is typically⇡ 16% of Sresc for proton initiated showers and 5% for ironinitiated showers.

srec contains i) the uncertainty in the reconstruction ofS(1000), ii) the uncertainty in Sresc due to the uncertaintyin the calorimetric energy measurement, and iii) the uncer-tainty in Sresc due to the uncertainty in Xmax; srec is typi-cally 12% of Sresc. ssim contains the uncertainty in Sresc dueto the uncertainty in S

µ

and SEM from the S(1000)�wµ

fitand to the limited statistics from having only three simu-lated events; ssim is typically 10% of Sresc for proton initi-ated showers and 4% for iron initated showers.

The resultant model of si, j is checked using the 59 events,of the 411, which are observed with two FD eyes whoseindividual reconstructions pass all required selection cutsfor this analysis. The variance in the Sresc of each eye iscompared to the model for the ensemble of events. Allthe contributions to si, j are present in this comparisonexcept for sshwr and the uncertainty in the reconstructedS(1000). The variance of Sresc in multi-eye events is wellrepresented by the estimated uncertainties using the model.In addition, the maximum-likelihood fit is also performedwhere sshwr is a free parameter rather than taken from the

0.4 0.6 0.8

1 1.2 1.4 1.6 1.8

2

0.7 0.8 0.9 1 1.1 1.2 1.3

R µ

RE

Systematic Uncert.QII-04 p

QII-04 MixedEPOS-LHC p

EPOS-LHC Mixed

Figure 5: The best-fit values of RE and Rµ

for QGSJET-II-04 and EPOS-LHC, for mixed and pure proton composi-tions. The ellipses show the one-sigma statistical uncertain-ties. The grey boxes show the estimated systematic uncer-tainties as described in the text; these will be refined in aforthcoming journal paper.

models; no significant difference is found between the valueof sshwr from the models, and that recovered when it is a fitparameter.

The results of the fit for RE and Rµ

are shown in Fig.5 and Table 1 for each HEG. The ellipses show the one-sigma statistical uncertainty region in the RE � R

µ

plane.The systematic uncertainties in the event reconstructionof Xmax, EFD and S(1000) are propagated through theanalysis by shifting the reconstructed central values by theirone-sigma systematic uncertainties; this is shown by thegrey rectangles.1 As a benchmark, the results for a purelyprotonic composition are given as well2.

The signal deficit is smallest (the best-fit Rµ

is the closestto unity) in the mixed composition case with EPOS. Asshown in Fig. 6, the primary difference between the groundsignals predicted by the two models is the size of the muonicsignal, which is ⇡15(20)% larger for EPOS-LHC thanQGSJET-II-04, in the pure proton (mixed composition)cases respectively. EPOS benefits more than QGSJET-IIwhen using a mixed composition because the mean primarymass determined from the Xmax data is larger in EPOS thanin QGSJET-II [20].

4 Discussion and SummaryIn this work, we have used hybrid showers of the PierreAuger Observatory to quantify the disparity between state-of-the-art hadronic interaction modeling and observed at-mospheric air showers of UHECRs. The most important ad-vance with respect to earlier versions of this analysis[21], inaddition to now having a much larger hybrid dataset and im-proved shower reconstruction, is the extension of the anal-

1. The values of ssim, srec and sshwr and the treatment of system-atic errors used here will be refined with higher statistics MonteCarlo simulations and using the updated Auger energy and Xmaxuncertainties, for the journal version of this analysis.

2. Respecting the observed Xmax distribution is essential for evalu-ating shower modeling discrepancies, since atmospheric attenu-ation depends on the distance-to-ground. This is automatic inthe present analysis, but the simulated LPs – which are selectedto match hybrid events – is a biased subset of all simulatedevents for a pure proton composition since with these HEGspure proton does not give the observed Xmax distribution.

54

Req

. res

calin

g of

µ-s

igna

l

Required rescaling of ground signals

The muon content of UHECR air showers observed by the Pierre Auger Observatory

The Pierre Auger Collaboration(Dated: January 17, 2014)

The hybrid events of the Pierre Auger Observatory are used to test the leading, LHC-tuned,hadronic interaction models. For each of 411 well-reconstructed hybrid events collected at theAuger Observatory with energy 1018.8

− 1019.2 eV, simulated events with a matching longitudinalprofile have been produced using QGSJET-II-04 and EPOS-LHC, for proton, He, N, and Fe pri-maries. The ground signals of simulated events have a factor 1.3-1.6 deficit of hadronically-producedmuons relative to observed showers, depending on which high energy event generator is used, andwhether the composition mix is chosen to reproduce the observed Xmax distribution or a pure protoncomposition is assumed. The analysis allows for a possible overall rescaling of the energy, howeverit proves not to be needed.

PACS numbers: Pierre Auger Observatory, ultra-high energy cosmic rays, muons, hadronic interactions

INTRODUCTION

The ground-level muonic component of ultra-high en-ergy (UHE) air showers is sensitive to hadronic parti-cle interactions at all stages in the air shower cascade,and to many properties of hadronic interactions suchas the multiplicity, elasticity, fraction of secondary pi-ons which are neutral, and the baryon-to-pion ratio [1].Air shower simulations rely upon hadronic event gener-ators (HEGs), such as QGSJET-II [2], EPOS [3], andSIBYLL [4]. The HEGs are tuned on accelerator exper-iments, but when applied to air showers they must beextrapolated to energies inaccessible to accelerators andto phase-space regions not well-covered by existing ac-celerator experiments. These extrapolations result in alarge spread in the predictions of the various HEGs forthe muon production in air showers [5].

The hybrid nature of the Pierre Auger Observatory,combining both fluorescence telescopes (FD) [6] and sur-face detector array (SD) [7], provides an excellent experi-mental setup for testing and constraining models of high-energy hadronic interactions. Thousands of air showershave been collected which have a reconstructed energyestimator in both the SD and FD. The measurement ofthe longitudinal profile (LP) constrains the shower devel-opment and thus the signal predicted for the SD, at theindividual event level.

PRODUCTION OF SIMULATED EVENTS

In the present study, we compare the observed groundsignal of individual hybrid events to the ground signalof simulated showers with matching LPs. The datawe use for this study are the 411 hybrid events with1018.8 < E < 1019.2 eV recorded between 1 January 2004and 31 December 2012 and satisfying the event qualityselection cuts in [8, 9]. This energy range is sufficient tohave adequate statistics while being small enough thatthe primary cosmic ray mass composition does not evolvesignificantly.

For each event in this data set we generate Monte Carlo

0

5

10

15

20

0 200 400 600 800 1000 1200

dE/d

X [P

eV/g

/cm

2 ]

Depth [g/cm2]

Energy: 8.8 ± 0.5 EeVZenith: 34.7 ± 0.4o

Xmax: 697 ± 7 g/cm2χ2/d.o.f. (p): 0.75χ2/d.o.f. (Fe): 0.76

DataQII−04 Fe

QII−04 p

1

10

100

500 1000 1500 2000

S [V

EM]

Radius [m]

DataQII−04 Fe

QII−04 p

FIG. 1. Top: The measured longitudinal profile of a typicalair shower with two of its matching simulated air showers, fora proton and an iron primary, simulated using QGSJET-II-04. Bottom: The observed and simulated ground signals forthe same event.

(MC) simulated events with a matching LP, as follows:• Generate a set of showers with the same geometry andenergy, until 12 of them have an Xmax value within the1-σ reconstruction uncertainty of the real event.• Among those 12 generated showers select, based on theχ2-fit, the 3 which best reproduce the observed longitu-dinal profile (LP).• For each of those 3 showers do a full detector simu-lation and generate SD signals for comparison with thedata. We produce three simulated showers to adequately

The muon content of UHECR air showers observed by the Pierre Auger Observatory

The Pierre Auger Collaboration(Dated: January 17, 2014)

The hybrid events of the Pierre Auger Observatory are used to test the leading, LHC-tuned,hadronic interaction models. For each of 411 well-reconstructed hybrid events collected at theAuger Observatory with energy 1018.8

− 1019.2 eV, simulated events with a matching longitudinalprofile have been produced using QGSJET-II-04 and EPOS-LHC, for proton, He, N, and Fe pri-maries. The ground signals of simulated events have a factor 1.3-1.6 deficit of hadronically-producedmuons relative to observed showers, depending on which high energy event generator is used, andwhether the composition mix is chosen to reproduce the observed Xmax distribution or a pure protoncomposition is assumed. The analysis allows for a possible overall rescaling of the energy, howeverit proves not to be needed.

PACS numbers: Pierre Auger Observatory, ultra-high energy cosmic rays, muons, hadronic interactions

INTRODUCTION

The ground-level muonic component of ultra-high en-ergy (UHE) air showers is sensitive to hadronic parti-cle interactions at all stages in the air shower cascade,and to many properties of hadronic interactions suchas the multiplicity, elasticity, fraction of secondary pi-ons which are neutral, and the baryon-to-pion ratio [1].Air shower simulations rely upon hadronic event gener-ators (HEGs), such as QGSJET-II [2], EPOS [3], andSIBYLL [4]. The HEGs are tuned on accelerator exper-iments, but when applied to air showers they must beextrapolated to energies inaccessible to accelerators andto phase-space regions not well-covered by existing ac-celerator experiments. These extrapolations result in alarge spread in the predictions of the various HEGs forthe muon production in air showers [5].

The hybrid nature of the Pierre Auger Observatory,combining both fluorescence telescopes (FD) [6] and sur-face detector array (SD) [7], provides an excellent experi-mental setup for testing and constraining models of high-energy hadronic interactions. Thousands of air showershave been collected which have a reconstructed energyestimator in both the SD and FD. The measurement ofthe longitudinal profile (LP) constrains the shower devel-opment and thus the signal predicted for the SD, at theindividual event level.

PRODUCTION OF SIMULATED EVENTS

In the present study, we compare the observed groundsignal of individual hybrid events to the ground signalof simulated showers with matching LPs. The datawe use for this study are the 411 hybrid events with1018.8 < E < 1019.2 eV recorded between 1 January 2004and 31 December 2012 and satisfying the event qualityselection cuts in [8, 9]. This energy range is sufficient tohave adequate statistics while being small enough thatthe primary cosmic ray mass composition does not evolvesignificantly.

For each event in this data set we generate Monte Carlo

0

5

10

15

20

0 200 400 600 800 1000 1200

dE/d

X [P

eV/g

/cm

2 ]

Depth [g/cm2]

Energy: 8.8 ± 0.5 EeVZenith: 34.7 ± 0.4o

Xmax: 697 ± 7 g/cm2χ2/d.o.f. (p): 0.75χ2/d.o.f. (Fe): 0.76

DataQII−04 Fe

QII−04 p

1

10

100

500 1000 1500 2000S

[VEM

]Radius [m]

DataQII−04 Fe

QII−04 p

FIG. 1. Top: The measured longitudinal profile of a typicalair shower with two of its matching simulated air showers, fora proton and an iron primary, simulated using QGSJET-II-04. Bottom: The observed and simulated ground signals forthe same event.

(MC) simulated events with a matching LP, as follows:• Generate a set of showers with the same geometry andenergy, until 12 of them have an Xmax value within the1-σ reconstruction uncertainty of the real event.• Among those 12 generated showers select, based on theχ2-fit, the 3 which best reproduce the observed longitu-dinal profile (LP).• For each of those 3 showers do a full detector simu-lation and generate SD signals for comparison with thedata. We produce three simulated showers to adequately

Distance from core [m]

data

Ready to be submi?ed

p-assumpt.

Fe-assumpt.


•Clear observation of flux suppression

• Strongest existing bounds on EeV ν and γ• Strongest existing bounds on large scale anisotropies

• First hints on directional correlations to nearby matter

• Increasingly heavier composition above ankle

• pp cross section at ~10*ELHC, LIV-bounds, ...

•muon deficit in models at highest energies

• geophysics (elfes, solar physics, aerosols...)

Major Achievementsin the first 6 years of operation

41


Science Goals of Auger Upgrade

43

1. Elucidate the origin of the flux suppression, i.e. GZK vs. maximum energy scenario- fundamental constraints on UHECR sources- galactic vs extragalactic origin- reliable prediction of GZK ν- and -γ fluxes

2. Search for a flux contribution of protons up to the highest energies at a level of ~ 10%- proton astronomy up to highest energies- prospects of future UHECR experiments

3. Study of extensive air showers and hadronic multiparticle production above √s=70 TeV- particle physics beyond man-made accelerators- derivation of constraints on new physics phenomena

measure composition intoflux suppression region...

... and do so event-by- event

do good muon counting}improve muon counting in surface detector array:factor of 10 in event statistics

]2 [g/cmmaxX600 700 800 900 1000

)m

axµ

(N10

log

8.4

8.5

8.6 p QGSjetII.04

He QGSjetII.04

N QGSjetII.04

Fe QGSjetII.04


Nµmax vs Xmax

44

Xmax (g/cm2)

lg (N

µ )

p

FeN

He

E = 5·1019 eVQGSJet II.04

Muons may even outperform Xmax

at highest energies !


Five different realisations under test in the field • segmented tank • scintillators or RPCs beneath tanks or buried in ground • scintillators on top

muonic signals

EM signals

Selection will be done this summerbased on performance, reliability, readiness, cost, risk

Different Upgrade Options under StudyNeed to improve on em/mu separation in EAS


UHECRBoost in understanding UHECRs

Auger data ➙ change of paradigm at GZK energies:seem to see maximum energy of cosmic accelerator(s)

Precise data and modelling required!UHECR ⇔ LHC: mutual benefits

The True High-Energy Frontier of Physicsmost stringent tests on LIV, Space-Time Structure...

Upgrades of present observatories and

Preparation for Next Generation Observatories in Space and at Ground

Exploring the Universe at the highest energies with the Pierre...

Documents

Transcript of Exploring the Universe at the highest energies with the Pierre...