Anisotropies at the Highest Energies
Angela V. Olinto
The University of Chicago
CMB vs UHECR Anisotropies
1. UHECRS are Charged,
should ‘point’ above ~ 10 EeV Z
RL

= kpc Z-1 (E / EeV) (B / G)-1
RL = Mpc Z-1 (E / EeV) (B / nG)-1
where
weak deflection
EeV=1018 eV
E > 1019eV
Halo B?
Milky way
B ~ G
strong deflection
E < 1018eV
kpc
10 kpc
Extra-galactic B?
B < nG
Ecm = PeV (Euhecr /ZeV)1/2 (2A)1/2
E-3.1
LHC
Ankle
(1 particle /km2 yr)
Tevatron
Extreme Accelerators
Emax > 1020 eV
E-2.7
RHIC
Non-thermal
Phenomena
32 orders of magnitude
2. Cosmic
Ray
Spectrum is
Power law
12 orders of magnitude
Swordy ‘02
3. ‘Pointing” UHECRS E ~ 100s TeV CM
knee
ankle
4. No known mechanism acceleration
nor clear Counterparts
10% efficiency
1020eV Protons
1020eV Iron
GRBs
LHC
Log B(G)
others
Log L(cm)
E = Ze B L
5. Extragalactic propagation GZK imprint
p+cmb +  p + 0
n++
Proton Horizon
~1020 eV
GZK Cutoff
Greisen, Zatsepin, Kuzmin
1966
Cosmologically Meaningful Termination
Attenuation length
e+e–

Attenuation length
Interaction length
6. GZK Effect  Anisotropies
 Spectral Feature
+ LSS
 Anisotropic Sky Distribution
 Charged Particle
Astronomy!!!
GZK Horizon - Protons
Allard, AO, Parizot 08
Horizons:
1019 eV ~ 1 Gpc
Gpc
100 Mpc
1020 eV < 100 Mpc
Inhomogeneous Galaxy Distribution
2MASS - Two Micron All Sky Survey
Horizons:
1019 eV ~ 1 Gpc
Gpc
100 Mpc
1020 eV < 100 Mpc
The Pierre Auger Observatory
of Ultra-High Energy Cosmic Rays
Northern site
20 000 km2
Argentina
Australia
Brasil
Bolivia*
Croatia*
Czech Rep.
France
Germany
Italy
Mexico
Netherlands
Poland
Portugal
Slovenia
Spain
UK
USA
Vietnam*
*Associate Countries
> 400 PhD scientists from
> 70 Institutions
and 18 countries
3 000 km2
Southern site
18 Countries
70 Institutions, > 400 scientists
Many Candidates to win the
World Cup 2010!!!
Auger South
1660 tanks = 3,000 km2
Surface Array
4 Fluorescence Detector
Sites
Science Data since 2004
Completed in June 2008
surface detector
tanks aligned seen from Los Leones
4 Fluorescence Detector Sites
1st - 4 Fold Hybrid Event
20 May 2007
E ~ 1019 eV
> ~1 degree “astronomy” @ 50 EeV
Statistical 14% @ 50 EeV
Systematic 22%
Less than 1 degree @ 50 EeV
Observed Spectrum showns
GZK GZK
Unclear Trend in Observed Composition
Anisotropy Analysis
Auger Anisotropy Result 2007
Highest energy cosmic rays
have an anisotropic distribution!
Data: 27 events above 57 EeV
from 1/1/2004 to 31/8/2007
(exposure of 8890 km2 sr y)
Correlation with z < 0.018 AGNs in
the 12th Véron-Cetty/Véron catalogue
The observed correlationdid not identify the CR sources
but established anisotropy at 99% c.l.
Auger 2007 Anisotropy Result
* - VCV AGN catalogue
3.2o- 27 highest energy event
Blue shades: equal integrated exposure
Challenges of the Analysis:
1. LOW STATISTICS!!!!!
2. Choice of catalogs
Véron-Cetty / Véron, 12th Edition, 2006
“This catalogue should not be used for any statistical
analysis as it is not complete in any sense, except that it
is, we hope, a complete survey of the literature.”
Significance of the anisotropy result
Not an “AGN correlation” result!
1st step: search for correlations between arrival directions of
UHECR and various source catalogues
(data from 2004/01/01 to 2006/05/27)
Very large “raw significance” found w/ 12th VCV catalogue of AGNs
Even after generous penalty factors for a posteriori searches
and scanning of parameter space
Did not seem to be fluctuation
Auger collaboration set up a prescription for future data
Most significant a posteriori “correlation signal”:
3.2 expected
from isotropic
distribution
12 out of 15 events E > 56 EeV closer than 3.1° from an
AGN in 12th VCV with z ≤ 0.018 (D ≤ 75 Mpc)
Exploratory scan (Period I): 1 Jan 2004-27 May 2006
Integrated exposure: 4390 km2 sr y
Scan in ψ(angular distance between CR and AGN), Eth (CR energy),
zmax (AGN distance)
Scan implemented to find the minimum of the binomial probability
P that ≥ k out of N events correlate by chance
p(z, ψ)=
exposure-weighted chance
probability for a CR to fall
within ψ of the sources
Minimum value of P: Eth = 56 EeV ψ = 3.1o zmax = 0.018
12/15 events correlate (3.2 expected by chance, p(z, ψ) = 0.21)
Scan  proper penalization difficult to calculate
 Prescription adopted
3 Parameter Scan
 ≤ 3.1°
z ≤ 0.018
(D ≤ 75 Mpc)
E ≥ 56 EeV
Energy Threshold makes sense
correlation is most significant above E = 56 EeV where the CR flux drops…
Significance of the anisotropy result
2nd step: predefine a region in the sky of excess CR flux &
test if next UHECRs come from this region
21% chance from isotropic distribution
Independent data set, prescribed parameters,
unambiguous significance
Nearby VCV AGN
21% sky
Independent data (Period II): 27/5/2006-31/8/2008
Integrated exposure: 4500 km2 sr y
A-priori fixed parameters: Eth = 56 EeV ψ = 3.1o zmax = 0.018
8/13 events correlate (2.7 expected by chance)
Probability to happen by chance from an isotropic flux: P≈1.7 10-3
Test built to have 1% probability to
incorrectly reject isotropy.
Test passed: 99% c.l. anisotropy
Note: it does not show that VCV AGN are UHECR sources!
First 27 events
Exploratory scan – 12/15 events correlated (3.2 expected)
Prescription passed when 8/13 correlated (2.7 expected)
NOTE: original set became 14 events when reconstructed with new energy analysis, 14+13 = 27 
References
Science 318 (2007), 939
“Correlation of the highest energy cosmic rays with nearby extragalactic objects”
Astroparticle Physics 29 (2008), 188
“Correlation of the highest-energy cosmic rays with the positions of nearby
active galactic nuclei”
arXiv:0906.2347
31st International Cosmic Ray Conference, Lodz , Poland, July 2009
“Astrophysical Sources of Cosmic Rays and Related Measurements with the
Pierre Auger Observatory”
NEW Update to come soon...
ICRC Update 2009
ICRC new data (Period III): 27/5/2006-31/3/2009
after exploratory period (II+III) integrated exposure: 12650 km2sr y
Updated event reconstruction: 56 EeV  55 EeV
31 new events E > 55 EeV: 44 post-prescription events
A-priori fixed parameters: Eth = 55 EeV ψ = 3.1º zmax = 0.018
mask = cut out
Galactic Plane
ICRC 2009
Likelihood ratio:
binomial probability of correlation
over binomial probability
in isotropic case (piso=0.21)
17/44 events in correlation (P=0.006)
Isotropy still rejected at 99% c.l.
Time Ordered Degree of correlation (p=k/N)
pdata=17/44=0.38±0.07
Correlation degree decreased wrt our previous report
but still more than 2 s.d. from isotropy
A Posteriori Analysis
Point source searches
Different methods & Catalogs (w/ ICRC 2009 Data):
I. Cross-correlation
II. Log-likelihood
III. 2-fold correlation
Laura Watson: Bayesian Methods applied to 2007 data
Fred Kuehn: Methos used in the collaboration
Centaurus A
Centaurus A
Integral number of events vs
angular distance from Cen A
KS test: 2% of isotropic realizations
have a maximum departure
from isotropy ≥ the maximum
observed departure
Maximum excess @ 18o radius: 12 observed events vs 2.7 expected
Other (Better) Catalogs
1. X-ray AGNs in 22-months SWIFT-BAT catalogue (261
objects)
2. Galaxies in the HIPASS survey (3058 galaxies, radio, fluxlimited, HI sources)
3. HIPASS-HL: sub-sample of HIPASS (the 765 most luminous
galaxies)
4. Galaxies in the 2MRS compilation (|b|>10¡, 23000 most
luminous galaxies from 2MASS, full-sky, IR)
Cross Correlation
Measures the excess of pairs within a given angular separation wrt
isotropy (departures are larger if CRs correlate with denser regions
of the catalogues)
e.g.: volume selected galaxies
(D<200 Mpc)
from 2MRS (1940 obj)
No weight by distance or luminosity (equal contribution to CR flux from every source)
Smoothed Maps from Catalogs
For each catalogue: Smoothed probability maps of arrival directions
Weight catalogue sources by flux x GZK attenuation
2 free parameters: smoothing angle σ and isotropic background (catalogue
incompleteness + lower en. spill)
Fc(n) = density map value
in the direction n
Use data to find the best values of
(σ, fiso) by
maximizing the LogLikelihood
Parameters weakly constrained
Likelihood Test
Simulate 104 samples with the same
number of data as in the real set
Draw events according to catalogue
density map or isotropically
Compare the likelihood distributions
with the value obtained from data
Better agreement with Catalog models
vs Isotropy
LL sensitive to data points “in or out” of
high density catalogue regions
Simulated data sets based on isotropy (I) and
Catalog models (II) compared to data (black line/point).
2-fold correlation method
Complementary method to test
overall proportionality between
models and data
Based on the computation of two
coefficients: correlation (test
match between models and data)
and concentration (test clustering
properties of data)
Data compatible with all the
models
The map based on SWIFT gives
the most discriminant test against
isotropy
Conclusions
up to 3/31/2009, exposure 17,000 km2sr.y, 58 CRs w E>55
EeV
Update on the correlation with VCV objects (a-priori analysis)
Isotropy rejected at 99% c.l.
Degree of correlation lower than in early data
A-posteriori analysis:
Excess of events close to Cen A (?)
Distribution of arrival directions compatible with several catalogs
that trace the distribution of nearby extra-galactic matter
X-rays AGNs (e.g., SWIFT-BAT) interesting correlations
More statistics needed to discriminate possible source scenarios
“Only” collect ≈ 2 UHECR/month
Auger North 10 yr
Black – 2MASS catalog z<0.02
Blue – Auger North simul Data – energy proportional to circle size
Red – Auger South Data
50
CERN,
— Particle
11 December
Physics2007
Seminar: Auger results on UHECRs —