Mass composition results from the Pierre Auger Observatory Simone - - PowerPoint PPT Presentation

Cosmic Ray International Seminar 13-17 September 2010 Catania - Italy

Mass composition results from the Pierre Auger Observatory

Simone Riggi

• n behalf of the Pierre Auger Collaboration

C.S.F.N.S.M. University of Catania INFN Catania

The role of composition measurements

Impact of composition information in CR understanding

1

Explanation of the spectral features → i.e. the nature of the knee, ankle and flux suppression above 5×1019 eV

2

Strong discrimination observable among different CR models → i.e. significant photon fluxes in top-down scenarios vs astrophysical scenarios

3

Support/cross-check to anisotropy analysis → benefit for correlation analysis with astrophysical objects Composition measurements rather uncertain above 1019 eV

2 / 43

The Pierre Auger Observatory

◮ Southern Observatory (Malargüe - Argentina): area ∼3000 km2

→ Hybrid design, completed in March 2008, in data taking since 2004 → Surface Detector (SD): 1660 water Cherenkov tanks → Fluorescence Detector (FD): 4 sites × 6 telescopes → Atmospheric monitoring devices: 4 Lidar stations, Central Laser Facility (CLF), weather stations, radio balloon soundings, IR cloud cameras, . . . → R&D activities: high elevation telescopes (HEAT), additional array with finer granularity (AMIGA), radio antennas and muon counters

◮ Northern Observatory (Colorado - USA): area ∼21000 km2 3 / 43

The Surface Detector (SD)

◮ 12 tonnes of deionized water ◮ 3 Photonis PMTs (diameter 12 cm) ◮ FADC sampling rate 40 MHz ◮ Solar panels for power supply ◮ Time tagging with GPS system (∼8 ns res-

• lution), antenna for data transfer

◮ Tank signal calibration in Vertical Equiva-

lent Muons (VEMs)

◮ 5 SD trigger levels, T1&T2 @ PMT and

tank level

◮ Tank event rate 3 kHz

trigger

− − − − → 3 events/day

4 / 43

SD Data Reconstruction

A real SD event (AugerId 200733602278) view from top sample ADC trace Event footprint: group of triggered tanks

◮ Signal-weighted barycenter of triggered tanks → core location ◮ Fit to tank timings with a shower front model → shower axis ◮ Fit LDF to tank signals ⇒ S1000 FD calibration

− − − − − − − − − − → Energy

5 / 43

SD Data Reconstruction

lateral profile Event footprint: group of triggered tanks

◮ Signal-weighted barycenter of triggered tanks → core location ◮ Fit to tank timings with a shower front model → shower axis ◮ Fit LDF to tank signals ⇒ S1000 FD calibration

− − − − − − − − − − → Energy

6 / 43

The Fluorescence Detector (FD)

◮ 6 telescopes/FD site ⇒ 24 telescopes; ◮ Telescope aperture: 30◦×30◦ ◮ Camera: 22×20 Photonis PMTs ◮ FADC sampling rate: 10 MHz ◮ Time tagging with GPS system ◮ Absolute and relative calibration with UV

LED sources

◮ Atmosphere calibrated with many devices

(Lidar, CLF, radio soundings, . . . )

◮ 3 FD trigger levels, FLT @ PMT level,

SLT&TLT @ camera level, T3 hybrid trigger @ FD level

◮ Hybrid event rate ∼ 5-10 events/hour 7 / 43

Composition studies in Auger

FLUORESCENCE DETECTOR

→ Xmax → RMS(Xmax)

HYBRID DETECTOR SURFACE DETECTOR

→ Signal rise time → Signal rise time asymmetry

8 / 43

Composition from hybrids

Depth of shower maximum Xmax

]

2

slant depth X [g/cm

200 300 400 500 600 700 800 900 1000 1100 )

9

Number of charged particles (x 10 1 2 3 4 5 6 7 8 eV

19

proton, E=10 eV

19

iron, E=10 eV

19

gamma, E=10

MC profiles for p, Fe and γ ]

2

[g/cm

max

X

500 600 700 800 900 1000 1100 1200

entries

50 100 150 200 250 300 350 400

Xmax distributions for p, Fe and γ

→ protons develop deeper in atmosphere and fluctuate more than nuclei → average Xmax and its fluctuations measured with great precision

9 / 43

Composition from hybrids: Xmax reconstruction

A real FD event (AugerId 200530502095)

3D view camera view

Event footprint: sequence of hit PMTs forming a track in the camera

◮ signal-weighted fit to PMT directions → Shower Detector Plane (SDP) ◮ signal-weighted fit to PMT timings → shower axis 10 / 43

Composition from hybrids: Xmax reconstruction

longitudinal profile Xmax reconstruction bias

How to reconstruct Xmax?

◮ light profile at the telescope aperture → energy deposit profile ◮ Gaisser-Hillas fit to profiles → Xmax

Xmax reconstruction systematics

◮ including rec algorithm, choice of longitudinal fitting function & lateral distribution ◮ systematic uncertainty < 8 g/cm2 (@1018 eV) 11 / 43

Composition from hybrids: event selection

Hybrid data from December 2004 - March 2009, Energy>1018 eV

◮ CALIBRATION SELECTION

→ no bad pixels

◮ ATMOSPHERE SELECTION

→ small cloud coverage and optimal aerosol conditions

◮ GEOMETRY SELECTION

→ dtank−axis < 2 km, θview>20◦ → precise measurement of the shower axis (∼0.1◦)

◮ PROFILE SELECTION

→ optimal GH fit, small Xmax uncertainties (<40 g/cm2) → no gaps in profiles → Xmax observed in field of view (FoV) → unbiased measurement of Xmax

∼1.5×106 raw hybrids selection − − − − − − − − → ∼ 3754 hybrids for physics analysis

12 / 43

Composition from hybrids: event selection

[E/eV]

10

log

18

10

19

10

Efficiency Ratio

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Fe

/Eff

p

Eff

Fe

/Eff

He

Eff

Fe

/Eff

O

Eff

Requiring Xmax in FOV reject very deep and shallow showers → bias the event selection towards light primaries How to avoid biases in selection? → Viewable depths depend on shower geometry (telescope FOV, distance from the FD), energy, atmosphere → Find a range of detectable depths on a event-by-event basis → Require Xmax within the fiducial boundary → Guarantee an unbiased selection in Xmax (<5 g/cm2) and RMS (<3 g/cm2)

13 / 43

Composition from hybrids: event selection

Good and bad events. . .

Good! Cloudy conditions! Xmax outside FoV! Bad aerosol conditions!

14 / 43

Composition from hybrids: Xmax resolution

total resolution resolution check with stereo data

◮ The resolution deteriorates towards low energies (less light) ◮ 1018 eV: ∼27 g/cm2, 1019 eV: <20 g/cm2 ◮ Cross-checked with stereo data 15 / 43

Nuclear composition from hybrids

Elongation rate → Change of composition around the ankle → Increase of the average mass up to 59 EeV → No zenith angle bias

16 / 43

Nuclear composition from hybrids

Elongation rate E/eV

18

10

19

10

]

2

> [g/cm

max

<X

600 650 700 750 800 850

QGSJET01 QGSJETII Sibyll 2.1 EPOS 1.99

Auger - PRL 2009 contour

syst

σ 1 ± HiRes - 2009

proton iron

→ AUGER and HIRES results compatible within systematics

17 / 43

Nuclear composition from hybrids

Xmax RMS → Increase of the average mass up to 59 EeV → No zenith angle bias

18 / 43

Nuclear composition from hybrids

Xmax RMS E/eV

18

10

19

10

]

• 2

RMS [g cm

max

X

10 20 30 40 50 60 70

→ AUGER and HIRES results compatible within systematics

19 / 43

Composition reconstruction with the SD

Rise time t1/2 in surface stations

t [25 ns]

60 65 70 75 80 85 90 95 100 105

Signal [VEM peak]

20 40 60 80 100 120

t [25 ns]

60 65 70 75 80 85 90 95 100 105

Signal [VEM peak]

20 40 60 80 100 120

±

µ γ

±

e total

◮ t1/2 sensitive to shower development

higher particle production heights (shallow showers) → narrow time pulses (smaller t1/2)

◮ t1/2 sensitive to electron/muon content

muons produce narrow pulses in tanks → muon-rich showers (nuclei) have smaller t1/2

◮ t1/2 linearly correlates with Xmax

benchmark ∆= 1 N

N

• i=1

ti

1/2 − t1/2(r, θ, Eref )

σi

1/2(r, θ, S)

20 / 43

Composition reconstruction with the SD

Asymmetry of rise time in surface stations

◮ em component absorption in late region

→ early-late asymmetry (dependence on azimuth ξ) → muons dominate in late regions → smaller t1/2 in late regions

◮ em absorption increases with zenith θ

→ muon component almost asymmetry-free → asymmetry decreases with θ

◮ Asymmetry profile maximum as

composition indicator

t1/2 r

• = a+bcos ξ

→ asymmetry profile b a (sec θ) → asymmetry maximum linearly correlates to Xmax

21 / 43

Composition reconstruction with the SD

Use hybrids to calibrate t1/2 & AsymmMax with Xmax t1/2 vs Xmax AsymmMax vs Xmax

22 / 43

Nuclear composition from SD

AsymmMax vs energy from SD data → Increase of the average mass with energy → Completely independent SD results!

23 / 43

Nuclear composition from SD

Average Xmax from SD data → Increase of the average mass with energy → Support FD results

24 / 43

Summary

Results from Auger using data collected during construction

◮ Independent measurement of composition with 2 detectors ◮ Parameter reconstruction and selection under control ◮ Different sources of systematics studied

→ Increase of the average nuclear mass with energy (both FD & SD!)

More to come soon on composition. . .

◮ increasing statistics ◮ composition from Xmax distributions ◮ low energy enhancements (→ 1017 eV) ◮ composition with muons counters and radio antennas 25 / 43

Backup slides

Xmax resolution

27 / 43

Xmax systematics

28 / 43

Hybrid data selection

Hybrid data selection

◮ PRE-SELECTION

→ hybrids (1 tank at least); → E>1019 eV (photon analysis), E>1018 eV (ER);

◮ GEOMETRY SELECTION

→ "HottestTank"-axis distance<2 km; → Minimum viewing angle>20◦;

◮ PROFILE SELECTION

→ Xmax observed; → optimal Gaisser-Hillas fit: χ2

GH/Ndf <2.5, χ2 GH/χ2 linear <0.9;

→ σstat (Xmax )< 40 g/cm2, σstat (E)/E < 20%; → no gaps in profile (ProfileGap/TrackLength< 20%;

◮ CALIBRATION SELECTION

→ good cal period; → no bad pixels;

◮ ATMOSPHERE SELECTION

→ Aerosol MieDB; → VAOD@3km ≤ 0.1; → Cloud Coverage ≤ 25%;

29 / 43

Hybrid data selection

Hybrid data selection

◮ PRE-SELECTION

→ hybrids (1 tank at least); → E>1019 eV (photon analysis), E>1018 eV (ER);

◮ GEOMETRY SELECTION

→ "HottestTank"-axis distance<2 km; → Minimum viewing angle>20◦;

◮ PROFILE SELECTION

→ Xmax observed; → optimal Gaisser-Hillas fit: χ2

GH/Ndf <2.5, χ2 GH/χ2 linear <0.9;

→ σstat (Xmax )< 40 g/cm2, σstat (E)/E < 20%; → no gaps in profile (ProfileGap/TrackLength< 20%;

◮ CALIBRATION SELECTION

→ good cal period; → no bad pixels;

◮ ATMOSPHERE SELECTION

→ Aerosol MieDB; → VAOD@3km ≤ 0.1; → Cloud Coverage ≤ 25%;

30 / 43

Hybrid data selection

Hybrid data selection

◮ PRE-SELECTION

→ hybrids (1 tank at least); → E>1019 eV (photon analysis), E>1018 eV (ER);

◮ GEOMETRY SELECTION

→ "HottestTank"-axis distance<2 km; → Minimum viewing angle>20◦;

◮ PROFILE SELECTION

→ Xmax observed; → optimal Gaisser-Hillas fit: χ2

GH/Ndf <2.5, χ2 GH/χ2 linear <0.9;

→ σstat (Xmax )< 40 g/cm2, σstat (E)/E < 20%; → no gaps in profile (ProfileGap/TrackLength< 20%;

◮ CALIBRATION SELECTION

→ good cal period; → no bad pixels;

◮ ATMOSPHERE SELECTION

→ Aerosol MieDB; → VAOD@3km ≤ 0.1; → Cloud Coverage ≤ 25%;

31 / 43

Event with maximum outside FOV

Hybrid data selection

Hybrid data selection

◮ PRE-SELECTION

→ hybrids (1 tank at least); → E>1019 eV (photon analysis), E>1018 eV (ER);

◮ GEOMETRY SELECTION

→ "HottestTank"-axis distance<2 km; → Minimum viewing angle>20◦;

◮ PROFILE SELECTION

→ Xmax observed; → optimal Gaisser-Hillas fit: χ2

GH/Ndf <2.5, χ2 GH/χ2 linear <0.9;

→ σstat (Xmax )< 40 g/cm2, σstat (E)/E < 20%; → no gaps in profile (ProfileGap/TrackLength< 20%;

◮ CALIBRATION SELECTION

→ good cal period; → no bad pixels;

◮ ATMOSPHERE SELECTION

→ Aerosol MieDB; → VAOD@3km ≤ 0.1; → Cloud Coverage ≤ 25%;

32 / 43

Event with bad pixels

Hybrid data selection

Hybrid data selection

◮ PRE-SELECTION

→ hybrids (1 tank at least); → E>1019 eV (photon analysis), E>1018 eV (ER);

◮ GEOMETRY SELECTION

→ "HottestTank"-axis distance<2 km; → Minimum viewing angle>20◦;

◮ PROFILE SELECTION

→ Xmax observed; → optimal Gaisser-Hillas fit: χ2

GH/Ndf <2.5, χ2 GH/χ2 linear <0.9;

→ σstat (Xmax )< 40 g/cm2, σstat (E)/E < 20%; → no gaps in profile (ProfileGap/TrackLength< 20%;

◮ CALIBRATION SELECTION

→ good cal period; → no bad pixels;

◮ ATMOSPHERE SELECTION

→ Aerosol MieDB; → VAOD@3km ≤ 0.1; → Cloud Coverage ≤ 25%;

33 / 43

Event with bad aerosol conditions

Hybrid data selection

Hybrid data selection

◮ PRE-SELECTION

→ hybrids (1 tank at least); → E>1019 eV (photon analysis), E>1018 eV (ER);

◮ GEOMETRY SELECTION

→ "HottestTank"-axis distance<2 km; → Minimum viewing angle>20◦;

◮ PROFILE SELECTION

→ Xmax observed; → optimal Gaisser-Hillas fit: χ2

GH/Ndf <2.5, χ2 GH/χ2 linear <0.9;

→ σstat (Xmax )< 40 g/cm2, σstat (E)/E < 20%; → no gaps in profile (ProfileGap/TrackLength< 20%;

◮ CALIBRATION SELECTION

→ good cal period; → no bad pixels;

◮ ATMOSPHERE SELECTION

→ Aerosol MieDB; → VAOD@3km ≤ 0.1; → Cloud Coverage ≤ 25%;

34 / 43

Event with cloudy conditions

Hybrid data selection

Hybrid data selection

◮ ANTIBIAS CUTS

→ requiring Xmax in FOV reject very deep and shallow showers → bias the event selection; → use MC to determine a range of detectable depths, given the shower geometry; → the viewable depths depend on shower geometry (telescope FOV, distance from the FD), energy, atmosphere; → guarantee an unbiased selection;

35 / 43

Hybrid data selection

Quality cuts applied to data Cut Events

• Rej. Factor [%]

Pre-Sel log10E[17.6,19.8],θ[0◦,60◦],hybrid,FdRecLevel=10 137561

• LL/LM/CO

116958 15.0 Cal Good Cal Period 112042 4.2 No Bad Pixels 109716 2.1 Atm Aerosol MieDB 91843 16.3 VAOD@3km ≤ 0.1 88643 3.5 Cloud Coverage ≤ 25% 65075 26.6 Geometry Rp > 0 65074 1.5·10−3 0<HotTankAxisDist<2000 m 64439 1.0 MinViewAngle > 20◦ 59295 8.0 Profile XTrackMin<Xmax<XTrackMax 35132 40.8 χ2

GH/Ndf < 2.5

34392 2.1 (χ2

linear-χ2 GH) > 4

15662 54.5 σ(Xmax)< 40 g/cm2 15454 1.3 ∆E/E < 20% 15454 0.0 ProfileGap/TrackLength< 20% 15307 1.0 Anti Bias Energy dependent HotTankAxisDist Cut 1 13888 9.3

1

HotTankAxisDist <    750 833 × log10(E/EeV) 2000 log10(E/EeV) < 0 0 ≤ log10(E/EeV) < 1.5 log10(E/EeV) ≥ 1.5

36 / 43

Definitions for Xmax and RMS

ER definition Xmax = 1 N

• j

X j

max

σ(Xmax) =

• j(X j

max − Xmax)2

N(N − 1) RMS definition V =

• j(X j

max − Xmax)2

(N − 1) RMS = √ V σ(RMS) =

• 1

N

• m4 − N−3

N−1 V 2

• √

4V RMScorr =

• V − σ(Xmax)2

σ(RMScorr) = √ V RMS σ(RMS)

37 / 43

Auger highlights in the field - Spectrum

Energy spectrum from SD, hybrid and inclined events

J.Abraham et al., Phys. Rev. Lett. 101 (2008), 061101 J.Abraham et al., to appear in Phys. Lett. B (2010)

→ Ankle observed at 1018.6 eV → Flux suppression above 4×1019 eV with 12σ significance

38 / 43

Auger highlights - Energy spectrum

Measurement of the energy spectrum J(E)= d4N dE dS dΩ dt ∼ 1 ∆E ND(E) ε(E)

◮ SD spectrum

⇒ zenith angles < 60◦; ⇒ energy > 3×1018 eV; ⇒ energy calibrated with a subset of high quality hybrid events; ⇒ exposure ε from number of active detector stations over time;

◮ FD spectrum

⇒ energy > 1018 eV; ⇒ exposure from dedicated time-dependent MC simulations describing real data-taking conditions;

39 / 43

Auger highlights - Energy spectrum

Combining SD & FD spectrum ⇒ Ankle observed at 1018.6 eV ⇒ Evidence for a flux suppression above 4×1019 eV with 6σ significance

40 / 43

Auger highlights in the field - Arrival Directions

Sky map of the most energetic events seen by Auger

J.Abraham et al., Science 318 (2007), 938 J.Abraham et al., Astropart. Phys. 29 (2008), 188 J.Abraham et al., arXiv:0906.2347 (2009)

→ Anisotropy and correlation with AGNs above 55 EeV → Born of cosmic ray astronomy and source identification?

41 / 43

Auger highlights - Anisotropy and correlation with AGNs

Correlation with VCV-catalogue (AGNs)

◮ Test null hypothesis (isotropy)

⇒ Prob. P to observe at least k correlated events out of N (isotropic p=0.21), likelihood ratio R: P =

N

• j=k

N j

• pj (1−p)N−j

R = 1

p pk (1 − p)N−k dp

pk (1 − p)N−k+1

◮ Find correlation scale

⇒ Scan P minima vs θ, z, Ethr

exploratory scan (01/2004-05/2006) ⇒ θ=3.1◦, zmax=0.018 (75 Mpc), Emin= 55 EeV, k/N=9/14 (kiso=2.9)

◮ Confirm/reject the signal with independent data samples

prescription: α=1%, β=5%, R>95% ⇒ reject isotropy @ 99% level

◮ Running the prescription

⇒ Data period II (06/2006-08/2007) k/N=9/13 (kiso=2.7), P=2×10−4 ⇒ ANISOTROPY CLAIM ⇒ Data period II+III (06/2006-03/2009) k/N=17/44 (kiso=9.2), P=6×10−3 ⇒ ANISOTROPY CLAIM. . . but reduced correlation signal

42 / 43

Auger highlights - Upper limit in photon flux

Upper limits of photon fraction

◮ SD limits

→ combination of t1/2-Rc as discrimination observable; → limits @ 10,20,40 EeV: 3.8,2.5,2.2×10−3 km−2sr−1yr−1;

◮ FD limits

→ Xmax as discrimination

• bservable;

→ limits @ 2,3,5,10 EeV: 3.8%, 2.4%, 3.5%, 11.7%; → Super-Heavy Dark Matter scenarios disfavored → No bias due to primary photons in energy calibration

43 / 43