TeV Particle Astrophysics 2010

Results from the HiRes Experiment

Gordon Thomson

University of Utah

Outline

• Introduction

• Some recent results– Spectrum– Composition – Anisotropy

• Summary

High Resolution Fly’s Eye (HiRes) Collaboration

J. Boyer, B. Connolly, C.B. Finley, B. Knapp, E.J. Mannel, A. O’Neill, M. Seman, S. Westerhoff

Columbia University

J.F. Amman, M.D. Cooper, C.M. Hoffman, M.H. Holzscheiter, C.A. Painter, J.S. Sarracino, G. Sinnis, T.N. Thompson, D. Tupa

Los Alamos National Laboratory

J. Belz, M. Kirn

University of Montana

J.A.J. Matthews, M. Roberts

University of New Mexico

D.R. Bergman, G. Hughes, D. Ivanov, S.R. Schnetzer, L. Scott, B.T. Stokes, S. Stratton, G.B. Thomson, A. Zech

Rutgers University

N. Manago, M. Sasaki

University of Tokyo

R.U. Abbasi, T. Abu-Zayyad, G. Archbold, K. Belov, J. Belz, D.R. Bergman, A. Blake, Z. Cao, W. Deng, W. Hanlon, P. Huentemeyer, C.C.H. Jui, E.C. Loh, K. Martens,J.N. Matthews, D. Rodriguez, J. Smith, P. Sokolsky, R.W. Springer, B.T. Stokes, J.R. Thomas, S.B. Thomas,

G.B. Thomson, L. Wiencke

University of Utah

The Two HiRes Fluorescence Detectors

• HiRes1: atop Five Mile Hill• 21 mirrors, 1 ring

(3<altitude<17 degrees). • Sample-and-hold electronics

(pulse height and trigger time).

• HiRes2: Atop Camel’s Back Ridge

• 12.6 km SW of HiRes1.• 42 mirrors, 2 rings

(3<altitude<31 degrees).• FADC electronics (100 ns

period).

Mirrors and Phototubes• 4.2 m2 spherical mirror

• 16 x 16 array of phototubes, .96 degree pixels.

Reconstruction • The trajectory of the EAS can be

determined in one of two ways:1. Monocular reconstruction using the

arrival time of light signal at the detector.

2. Stereo by intersecting the shower-detector planes (SDP) seen from the two detector sites. 2

tan0i

i c

Rtt P

i

0t

it

2.)

1.)

Spectrum and Composition Method

• Spectrum: aperture varies with energy. Must calculate with Monte Carlo method.– Trigger and thresholds very important.– Excellent accuracy is achievable.

• Composition: Monte Carlo simulations are essential to understand Xmax studies.

Compare Data to Monte Carlo: Judge success of simulation and

acceptance calculation.

Inputs to Monte Carlo: Fly’s Eye stereo spectrum; HiRes/Mia and HiRes Stereo composition; Library of Corsika showers.Detailed nightly information on trigger logic and thresholds, live mirrors, etc.

Result: excellent simulation of the data, and an accurate aperture calculation.

5σ Observation of the Break in the Spectrum

• Broken Power Law Fits– Two BP with extension

to test hypothesis that a break is present.

• Expect 43 events

• Observe 13 events

• Poisson probability: P(13;43) = 7x10-8,

which is 5.3

– The break is statistically significant.

Break is at (5.6 ± 0.5) x 1019 eV;GZK expected between 5 and 6 x 1019 eV.

The break is the true GZK cutoff.

E-5.1

Use Berezinsky’s E½ Method to Test

• E½ is the energy where the integral spectrum falls below the power-law extension by a factor of 2.

• Berezinsky et al.: log10E½ = 19.72, for a wide range of spectral slopes.

• Use 2 Break Point Fit with Extension for the comparison.

• log10E½ = 19.73 ± 0.07

• Passes the test.

Local Density of Sources

• Theoretical predictions for spectrum shape agree with HiRes measurements.

• Compare HiRes spectrum slope above the GZK energy to Berezinsky et al. predictions:– Line 1: constant density.– Line 5: no sources within

10 Mpc.– Line 2: double density

within 30 Mpc.

Berezinsky, Gazizov, and Grigorieva,Phys. Rev. D74, 043005 (2006) (uses older HiRes spectrum)

Local Density of Sources

• Compare HiRes spectrum slope above the GZK energy to Berezinsky et al. predictions:

– Line 1: constant density.– Line 5: no sources within 10

Mpc.– Line 2: double density within

30 Mpc.– Line 3: triple density within

30 Mpc.– HiRes: E-5.1 fall-off.

• More work is needed to make a better comparison, but...

• Constant density of sources is favored.

Berezinsky, Gazizov, and Grigorieva,Phys. Rev. D74, 043005 (2006) (uses older HiRes spectrum)

E-5.1

HiRes/Auger Spectra Comparison

Auger confirms all spectral features of HiRes spectrum

Spectrum Summary

• The GZK cutoff is present. The first observation was by the HiRes experiment.

• All details of the spectrum indicate the composition is protons.– The energy of the GZK cutoff is as expected for protons.– Highest energy extragalactic cosmic rays travel > 50 Mpc.– The fall-off above the cutoff is evidence for a constant density of

sources. CR’s travel a long distance. Spallation breaks up all nuclei at high energies proton flux results.

– The ankle has been observed by HiRes, at 1018.65 eV. The spectral index changes from -3.2 to -2.8

– Shape and energy of the ankle are consistent with e+e- production in collisions between extragalactic protons and photons of the CMBR.

More Direct Composition Measurement: use <Xmax>

• HiRes result on <Xmax>, and on σ(Xmax), indicate light composition.

• <Xmax> dependence is logarithmic: halfway is Be. Spallation at high energies breaks up light elements.

Composition is essentially all protons.

QGSJetII Protons Look Like Our Data(points=data, histogram=MC)

MC protons

MC iron

Data (points) vs. QGSJet-II MC (histogram)

Energy Bins: 18.3, 18.5

Energy Bins: 18.7, 18.9

Energy Bins: 19.1, 19.3

Energy Bins: 19.5, 19.7

Composition Summary

• The most direct indicator of composition is <Xmax>.

• Our <Xmax> data indicate a light composition, and favor protons.

• This is consistent with all of our spectrum information.

The Search for Anisotropy:

• Aim is to identify UHECR sources.• Two methods:

– Make a sky plot and look for bumps.

– Look for correlations with known astronomical object types.

• All cosmic ray results are of marginal significance:– Sky plots: AGASA doublets/triplet at 40 EeV

– Correlations: BL Lac’s at 10 EeV & AGN’s at 57 EeV

AGN Correlations

• Auger events’ correlations with AGN’s (south)– Early data set: scanned in (Emin, θ, zmax) using Veron-

Cetty+Veron catalog, found best correlations at (57 EeV, 3.1°, .018).

– Tested correlation with later data set, found 8/13 events correlated, chance probability of 0.002 (2.9σ).

• Not an “observation”: PRL requires 5σ. Require confirmation.

• Auger test using later data: 42 events, 12 correlated, expect 8.8 random, 1σ.

no effect.

HiRes Test of AGN Correlations (north)

• Choose Auger optimum point, same catalog.– 2/13 events correlated (expect 3 randomly)

– chance probability of 0.23 no effect.

• Scan the HiRes data – best point is (15.8 EeV, 2.5°, 0.016), 46/198

correlated.

– chance probability of 0.29 no effect.

Test of Correlations with Local Large Scale Structure

• All results extragalactic, point source; energies of 10 (BL Lac), 40 (clusters), 57 EeV (AGN).

• Matter is not distributed uniformly within the horizon of 57 EeV protons

Search for correlations with local large scale structure for high energy events with relatively close horizons. A priori choice: 10, 40, 57 EeV. Also choose, a priori, to quote 95% CL.

HiRes Stereo Data

• HiRes complete stereo data set, angular resolution ~0.8°

• Events not within 10° of galactic plane:– 10 above 57 EeV– 27 above 40 EeV– 310 above 10 EeV

• 6636 events in all

Local LSS Model• Based on 2 Micron All-Sky

Redshift Survey (2MRS)1, a flux-limited sample of galaxies with m ≤ 11.25– Remove galactic plane (|b| < 10°)

and objects within 5 Mpc.– Transform to a volume-limited

sample by weighting. – Result is 15,508 galaxies

between 5 and 250 Mpc.– Assume distribution is isotropic

beyond 250 Mpc.

1. J. Huchra, L. Macri, T. Jarrett, et al., in preparation.

Procedure: set of MC events coming from LSS

• Start with local LSS model.• Modify using HiRes aperture.• Simulate the data set; find average

predicted event density.• Two parameters:

– Minimum energy– Angular smearing to simulate

magnetic fields• Expect ~1°, extragalactic fields, for

E≥40 EeV1

• Expect 2°-4°, galactic fields. • Perform K-S test between data and

expectation from LSS.

• Repeat starting with an isotropic galaxy distribution.

1. T. Kashti and E. Waxman, JCAP 0805, 006 (2008).

57 EeV

40 EeV

10 EeV

Smearing angle of 6°

Sky Plots10 EeV

40 EeV

57 EeV

Results

• Choose 95% c.l. exclusion to quote, a priori.

• For isotropic model, get good agreement.

• For local LSS model get poor agreement.

• Exclude correlation at 95% c.l. for θs < 10°, E ≥ 40 EeV

LSS Conclusions • We have searched for correlations between the pointing

directions of HiRes stereo high energy events and local large scale structure.

• There are none at the 95% confidence level for magnetic smearing angles < 10°.

• This is very surprising.– One expects to see correlations.– One expects magnetic field smearing at the 4°-5° level.– With limited statistics we are able to place very significant limits.– Are the sources not in galaxies? Are magnetic field estimates wrong?

• This result should be confirmed by a future experiment (north).• The Telescope Array Experiment is now collecting data. It has

statistics equal to HiRes stereo (see TA talk, Yuichiro Tameda).

Summary• HiRes made the first observation of the GZK

cutoff. – It occurs at (5.6 ± 0.5 ± 0.9) x 1019 eV.

– Flux α E-5.1 ± 0.7 above the cutoff.

• We see the “ankle” at 1018.6 eV.– E-3.3 below and E-2.8 above.

• The composition is very light, probably protons• No AGN correlations are observed.• Correlations with local large scale structure are

very weak; magnetic fields may be larger than previously thought.