Search for clustering of ultra high energy cosmic rays from the Pierre Auger Observatory

0.19 MB·English

by  Silvia Mollerach

9
0
Search for clustering of ultra high energy cosmic rays from the Pierre
0
2
Auger Observatory

n
a Silvia Mollerach, for the Pierre Auger Collaboration a∗
J

9 aPierre Auger Observatory,av. San Mart´ın Norte 304, (5613) Malargu¨e, Argentina
2 CONICET, Centro Ato´mico Bariloche, 8400 Bariloche, Rio Negro, Argentina

]
E Wepresenttheresultsofasearchforclusteringamongthehighestenergyeventsdetectedbythesurfacedetector
H of the Pierre Auger Observatory between 1 January 2004 and 31 August 2007. We analyse the autocorrelation
function,inwhichthenumberofpairswithangularseparationwithinanangleαiscomparedwiththeexpectation
.
h from an isotropic distribution. Performing a scan in energy above 30EeV and in angles α < 30◦ , the most
p significant excess of pairs appears for E > 57EeV and for a wide range of separation angles, 9◦ <α<22◦. An
- excess like thishas a chance probability of ∼2% to arise from an isotropic distribution and appears at the same
o
r energy threshold at which the Pierre Auger Observatory has reported a correlation of the arrival directions of
t cosmic rays with nearby astrophysical objects.
s
a
[

1
v 1. Introduction sources (closer than ∼200 Mpc).
9 Theseideashavemotivatedanextensivesearch
9 The identification of the sources of the ultra
for clustering signals at high energies looking for
6 highenergycosmicrayshasbeenoneofthemain
excesses in the number of pairs at different an-
4
open problems in astrophysics since their discov-
. gular scales. Small angular scale searches essen-
1 ery. The study of their arrival directions is likely
tially look for cosmic rays coming from the same
0 to provide significant insight into this question.
9 source which would be the expected signal from
Ifcosmic raysare chargedparticles,protonsor
0 thestrongestnearbysources. Atintermediatean-
heavier nuclei, their trajectories are expected to
: gular scales, a clustering signal would be an evi-
v
be bent by the intervening galactic and extra-
i dence of the pattern characterising the distribu-
X galactic magnetic fields, and their arrival direc-
tion of the nearby sources, with pairs of events
tions will not point back to their sources. The
r resulting from cosmic rays coming from differ-
a intensity and orientation of these fields are not
ent sources. The angular scale up to which we
well known, but as the deflections are inversely
can expect that pairs of events come from the
proportional to the energy, the effect is smaller
same sourcedepends onthe energyofthe events,
at the largest energies. Thus, it is at the high-
onthe(unknown)magnitudeofinterveningmag-
estenergiesthat cosmic rayarrivaldirections are
netic fields, and on the cosmic ray composition.
most likely to trace their sources.
Although the data from a number of experi-
Ontheotherhand,thedistancefromwhichul-
ments have shown a remarkably isotropic distri-
trahighenergyprotonscanarrivetotheEarthis
butionofarrivaldirections,therehasbeenaclaim
expectedtobelimitedbytheenergylossescaused
of small scale clustering at energies larger than
bythe photo-pionproductionprocessesinthein-
40EeVbytheAGASAexperiment[2]. Themost
teraction with the cosmic microwavebackground
recent published analysis [3] reports 8 pairs (five
(Greisen Zatsepin Kuzmin, GZK effect [1]), and
doublets and a triplet) with separation smaller
similarlynuclei canundergophoto-disintegration than2.5◦ amongthe59eventswithenergyabove
processes. Hence, at energies above ∼ 60 EeV,
40EeV,while1.7wereexpectedfromanisotropic
cosmic rays should mostly come from nearby
flux. Theprobabilityforthisexcesstohappenby
∗[email protected]. chance was estimated to be less than 10−4. The
1
2
significance of the AGASA clustering result was, 2. The observatory and the data set
however, subject of debate based on the concern
The Pierre Auger Southern Observatory is lo-
thatthe energythresholdandangularseparation
cated in the Province of Mendoza, Argentina, at
were not fixed a priori. Tinyakov and Tkachev
◦ ◦ ◦ ◦
35.1 –35.5 S, 69.0 –69.6 W and 1300–1400 m
[4] computed the penalisation arising from mak-
a.s.l. The growing observatory has been in oper-
ingascanintheenergythresholdandobtaineda
probabilityof3×10−4. FinleyandWesterhoff[5] ationsince 2004: the datapresentedherereferto
theperiodbetween1January2004and31August
took alsointo accountthe penalisationfor a scan
2007,duringwhichthenumberofsurfacestations
intheangularscaleandobtainedaprobabilityof
3.5×10−3. The HiRes observatory has found no was increasing from154to 1388. The surface de-
tector consistsof10m2×1.2 mwaterCherenkov
significant clustering signal at any angular scale
◦ tanks spacedby 1500m, coveringanareathat in
up to 5 for any energy threshold above 10 EeV
◦ theperiodofinterestrangedfromabout200km2
[6]. A hint of correlation at scales around 25
to around 3000 km2. While the experiment has
and energies above40 EeV,combining data from
been described in detail elsewhere [13], the rele-
HiRes stereo, AGASA, Yakutsk and SUGAR ex-
vant featureswith respectto the presentanalysis
periments has been pointed out in Refs. [7,8].
will be outlined here.
ThePierreAugerObservatory,whoseconstruc-
The trigger requirement for the array is based
tion has recently been completed in Argentina,
on a 3-fold coincidence, that is satisfied when
has been taking data since January 2004. Its in-
a triangle of neighbouring stations is triggered.
tegratedexposure,up to the end ofAugust2007,
is 9000 km2 yr sr, representing the largest one The 50%contaminationfromaccidentaleventsis
obviated by an appropriate event selection [14],
ever attained by an extensive air shower arrayat
whose power for selecting real showers is larger
ultra high energies. Thanks also to its angular
than 99%.
and energy resolutions it offers an excellent data
The arrival directions of the showers are ob-
set forstudies ofanisotropiesin the arrivaldirec-
tained through the time of flight differences
tions.
among the triggered stations. The angular res-
Weapplyheretheautocorrelationtechniqueto
olution, defined as the angularradius aroundthe
the events with energy above 30 EeV, scanning
truecosmicraydirectionthatwouldcontain68%
overenergythresholdandangularseparation(up
◦ of the reconstructed shower directions, is calcu-
to 30 ), with the aim of searching for possible
lated on an event by event basis and checked
clusteringinthearrivaldirections. Apreliminary
throughthecorrelationswiththefluorescencede-
analysis has been presented in Ref. [9].
tector. Itdependsonthenumberoftriggeredsta-
The most clear anisotropy signal of the arrival
tions andis better than2◦ for3-foldevents(E <
directions of high energy events (with E > 57
◦
4 EeV), better than 1.2 for 4-folds and 5-folds
EeV) has been reported by the Auger collabo-
events (3 < E < 10 EeV) and better than 0.9◦
ration [10,11] through a completely independent
for higher multiplicity events (E > 10 EeV) [15].
method by studying their correlation with the
The estimatorfortheprimaryenergyisthe re-
nearbyextragalacticmatterdistribution. Theob-
constructedsignalat1000mfromtheshowercore
served correlation with nearby AGNs from the
[16]. The conversion from this estimator to en-
V´eron-Cetty and V´eron catalog [12] was shown
ergy is derivedexperimentallythroughthe use of
to be incompatible with an isotropic distribution
a subset of showers detected by both the surface
at more than 99% CL. The autocorrelationanal-
and the fluorescence detectors [16]. The energy
ysis discussed here does not depend on an a pri-
resolution is 18% and the absolute energy scale
ori selection of the possible sources location and,
has a systematic uncertainty of 22% [17].
therefore, gives complementary information with
In the present analysis, we apply the following
respect to that analysis.
cuts to the events:
• zenith angle θ <60◦;
3
• core location within the array boundaries: Thefactthatthedeflectionsexpectedfromgalac-
reconstructed core within a triangle of ac- tic and extragalactic magnetic fields and the
tive stations and station with the highest distribution of the sources are largely unknown
signalsurroundedbyatleast5activetanks; prevents us from fixing these values a priori.
The significance of an autocorrelation signal at
• reconstructed energy E >30 EeV.
a givenangle andenergy,when these values have
After these cuts there are 203 events above 30 not been fixed a priori, is a delicate issue that
EeVand81eventsabove40EeV.Fortheseevents has made, for example, the significance of the
the surface detector trigger efficiency is 100%. AGASA small scale clustering claim very con-
The acceptance is fully saturated and is deter- troversial. We adopt here the method proposed
mined by purely geometrical considerations [18], by Finley and Westerhoff [5], in which a scan
thus allowing an accurate calculation of the ex- over the energy threshold and the angular sep-
posure even with a changing array configuration. aration is performed. For each value of E and
2
Theexposureis flatasafunctionofsin θ, where α, we compute the fraction f of simulations hav-
θ isthezenithangle,anditisnearlyuniformasa ing an equal or larger number of pairs than the
functionofthe azimuthangleφ. There is a small data by generating 106 simulated isotropic data
modulation of the flux in right ascension due to sets modulated by the exposure. The fraction f
the dead times and the growth of the array dur- takes here the role of the distance between the
ingthedatatakingperiod,butthiseffectissmall, experimental and expected distributions as used
below1%,andcanbeignoredinthepresentanal- in the Kolmogorov-Smirnov test. The most rel-
ysis. evant clustering signal corresponds to the values
of α and E that have the smallest value of f, re-
3. The autocorrelation function analysis ferredheretoasf . Toestablishthestatistical
min
significanceofthisdeviation,itisnecessarytoac-
3.1. The method
count for the fact that the angular bins, as well
Astandardtoolforstudyinganisotropiesisthe
as the energy ones, are not independent. To do
two-pointangularcorrelationfunction. Inthefol- that, we perform 105 isotropic simulations with
lowing we will use the correlation function given
thesamenumberofeventsasthedataandcalcu-
by the number of pairs separatedby less than an
late for each realisation the most significant de-
angle α among the N events with energy larger
viation fi . The statistical significance of the
than a given threshold E, min
deviationfromisotropyis the integralofthe nor-
N i−1 malised f distribution above fdata. We can
min min
np(α)=XXΘ(α−αij), (1) then estimate the probability that such cluster-
i=2j=1 ing arises by chance from an isotropic distribu-
where α is the angular separation between tion just from the fraction of simulations having
ij
eventsiandj andΘisthestepfunction. Theex- fi ≤fdata.
min min
pectednumberofpairsandthe90%CLerrorbars
are obtained by generating a large number (106) 3.2. Results
of Monte Carlo simulations with the same num- Tostudytheautocorrelationfunctionusingthe
ber of events as in the real data set, isotropically scantechniquewehavetofixtherangeofthepa-
distributedandmodulatedbytheexposureofthe rameter space that we explore. We scan up to
◦
detector. The chance probability for any excess 30 in the angular scale, what includes the range
of pairs at a fixed angle α and energy threshold where small scale signals from point sources and
is found from the fraction of simulations with a intermediateangularscalesignalsfromclustering
largerorequalnumberofpairsthanwhatisfound ofthesourcesareexpectedtoappear. Wescanon
in the data at the angular scale of interest. energiesabove30EeV:thisrangeprobesthehigh
Theresultoftheautocorrelationfunctionanal- energy region where anisotropies are expected in
ysis depends on the chosen values of α and E. theastrophysicalsourcesscenario,andgoesdown
4
60
50
40
p 30
n
20
10
0
0 5 10 15 20 25 30
α
1
0.1
—
f 0.01
Figure 1. Autocorrelation scan for events with 0.001
◦
energy above 30 EeV and angles up to 30 . The
fraction ofsimulations with more pairs than that
1e-04
observed in the data is plotted for each angular 0 5 10 15 20 25 30
scale and number of events (or energy). α
Figure2. Upperpanel: Autocorrelationfunction
enough to cover the energy range where past
for the 27 events with energy largerthan 57 EeV
anisotropyclaimswithdatafrompreviousexper-
as a function of the angle (dots) and autocorre-
iments using a similar technique were made, al-
lation function for an isotropic distribution with
lowingforapossibleenergycalibrationdifference
90% confidence level band. Lower panel: Frac-
ofupto30%betweentheexperiments. Wethere-
tion of isotropic simulations with larger or equal
forecomputethenumberofpairsasafunctionof
number of pairs than the data.
◦ ◦ ◦
theseparationanglefrom1 to30 instepsof1 ,
and as a function of the number of events start-
ing fromthe10highestenergyeventsandadding
events one by one in decreasing energy order up for the 27 highest energy events (corresponding
to 100 events, and in groups of ten up to 200 to E > 57 EeV), for α = 11◦ (18 observed pairs
events, correspondingto a thresholdenergyof30 against 5.2 expected).
EeV, and then compare them with isotropic dis- InFigure2theautocorrelationfunction(upper
tributed simulations. panel) and the fraction of isotropic simulations
In figure 1 we present the fraction of simula- withequalorlargernumberofpairsthanthedata
tions withmorepairsthanthe dataasa function (lower panel) is shown for these 27 events as a
of the angular scale α and the number of events function of the angular scale. The broad region
N (or equivalently the energy threshold). of low values of f (< 10−3) is visible between 9◦
◦ ◦
Abroadregionwith an excessof pairsappears and 22 , with the minimum at 11 .
for energies above around 50 EeV, at angular The chance probability of a f ≤1.5×10−4
min
◦ ◦
scalesbetweenabout9 to22 . Inparticular,the to arise from an isotropic distribution was esti-
minimum value off (f ≃1.5×10−4)is found mated by performing the same scan to 105 simu-
min
5
tions showing a larger or equal number of pairs).
Duetoapossibledifferenceintheenergycalibra-
tion between AGASA and Auger, the clustering
signalreportedbyAGASAcouldappearinAuger
data at a lower energy scale. For energies above
30 EeV the observednumber of pairs is 23, while
the isotropic expectation is 16 (f = 0.05). This
excess is much smaller than the one reported by
AGASA.
4. Conclusions
Using the events recorded by the surface de-
tector of the Pierre Auger Observatory between
January2004andAugist2007wehaveperformed
a scan to search for possible clustering using the
autocorrelation technique. The scan has been
performed in energy (above 30 EeV, 203 events
◦
detected) and angular separation (between 1
◦
and 30 ). The most significant excess of pairs
is found for E > 57 EeV (corresponding to 27
Figure 3. Arrival directions of the 27 events events) and α = 11◦, with a chance probabil-
with energy larger than 57 EeV in equatorial co- ity of P = 1.6×10−2 to arise from an isotropic
ordinates centered in the south pole. The circles distribution. Above this energy we observe, in
aroundeacheventhavea5.5◦radius. Thedashed fact, a broad region of low probabilities, for an-
linerepresentsthe galacticplane (withthe galac- gularscalesbetween9◦and22◦. Athigherenergy
tic center identified by a cross)and the solid line thresholdsthechanceprobabilitiesbecomelarger,
represents the supergalactic plane. howeverthenumberofeventsbecomesratherlim-
ited. At lower energies the autocorrelation func-
tionbecomesprogressivelyclosertothatexpected
from an isotropic flux.
lations, and it is P ≃1.6×10−2. It should be noted that the energy threshold
Figure 3 shows the arrival directions of the 27 that maximises the significance is the same one
events with E > 57 EeV in equatorial coordi- as the energy at which the Pierre Auger collab-
natescenteredinthesouthpole. Thegalacticand oration has observed a departure from isotropy
supergalactic planes are displayed by the dashed using the correlation of the arrival directions of
andsolidlinesrespectively. Circlesof5.5◦around events with nearby AGNs. It is also the energy
each event guide the eye to identify pairs sepa- at which the cosmic ray flux at ultra high ener-
◦
ratedbylessthan11 . Thearrivaldirectionsand gies reported by the Pierre Auger collaboration
energies of the 27 events are listed in the Ap- [16] is suppressed by 50% compared to a power
pendix of [11]. law extrapolation of the flux measured at lower
The analysis of the autocorrelation function energies. This suppression, if interpreted as due
shows in particular no significant clustering sig- to the GZK horizon[1], wouldimply that nearby
nal at the small angular scales corresponding to sources dominate the flux at these energies. As
the AGASA claim. At energies above 40 EeV, the matter distribution in the nearby Universe
with81detectedevents,weobserve4pairswithin becomes increasingly anisotropic for smaller dis-
2.5◦, while 2.5 were expected from an isotropic tances, it is to be expected that the ultra high
flux (witha fractionf =0.24ofisotropicsimula- energy cosmic raysky shouldbecome anisotropic
6
attheseenergies,reflectingthepatternofthedis- Proc. 29th ICRC, Pune, India, 7 (2005) 71.
tribution of the sources.
REFERENCES
1. K. Greisen, Phys. Rev. Lett. 16 (1966) 748;
G. T. Zatsepin and V. A. Kuzmin, JETP
Lett. 4 (1966) 78.
2. N. Hayashida et al. [AGASA Collaboration],
Phys. Rev. Lett. 77 (1996) 1000.
3. M. Teshima et al. [AGASA Collaboration],
Proc 28th ICRC, Tsukuba, Japan (2003),
437.
4. P.G.TinyakovandI.I.Tkachev,JETPLett.
74 (2001) 1 [Pisma Zh. Eksp. Teor. Fiz. 74
(2001) 3], arXiv:astro-ph/0102101.
5. C. B. Finley and S. Westerhoff, Astropart.
Phys. 21 (2004) 359.
6. R.U.Abbasietal.[TheHighResolutionFly’s
Eye Collaboration (HIRES)], Astrophys. J.
610, L73 (2004).
7. M.KachelriessandD.V.Semikoz,Astropart.
Phys. 26, 10 (2006).
8. A. Cuoco, G. Miele and P. D. Serpico, Phys.
Rev.D74(2006)123008;A.Cuoco,G.Miele
and P. D. Serpico, Phys. Lett. B 660 (2008)
307.
9. S. Mollerach et al. [Pierre Auger Collabo-
ration], Proc. 30th ICRC, M´erida, Mexico
(2007), arXiv:0706.1749.
10. Pierre Auger Collaboration, Science 318
(2007) 938.
11. PierreAugerCollaboration,Astropart.Phys.
29, 188 (2008).
12. M.-P.V´eron-CettyandP.V´eron,Astron.and
Astrophys. 455 (2006) 773.
13. J. Abraham et al. [Pierre Auger Collabora-
tion], Nucl. Inst. and Meth. A523 (2004)50.
14. D. Allard et al. [Pierre Auger Collaboration],
Proc. 29th ICRC, Pune, India, 7 (2005) 287.
15. C. Bonifazi et al. [Pierre Auger Collabora-
tion], these proceedings.
16. J. Abraham et al. [Pierre Auger Collabora-
tion], Phys. Rev. Lett. 100, 061101 (2008).
17. B. Dawson et al. [Pierre Auger Collabo-
ration], Proc. 30th ICRC, M´erida, Mexico
(2007), arXiv:0706.1105.
18. D. Allard et al. [Pierre Auger Collaboration],