the Pierre Auger Observatory Cosmic Rays of UltraHigh

the Pierre Auger Observatory (Cosmic Rays of Ultra-High Energy) • • • The puzzle of UHECR Principle and advantages of an hybrid detector Present status of the Observatory Sensitivity to hadronic modelling at UHE Perspectives Pierre Billoir, LPNHE (CNRS/univ. Paris 6/Paris 7) (Auger Collaboration) EDS 05, Blois, may 2005

Energy Spectrum of Cosmic Rays Galactic ? Extragalactic ? ankle knee ? ?

Open issue: the end of the spectrum Below GZK: AGASA (ground array) and Hi. Res (fluorescence) almost agree (30 % systematic on E ? ) Above GZK: unexplained divergence ! AUGER has to find the truth…with an hybrid detector GZK cutoff after Douglas Bergman

Modelling of shower development (1) • hadronic cascade: ~1/3 lost at each step (p 0 2 g) ~2/3 re-interacts until decaying into muons (E ~ a few Ge. V) • electromagnetic cascade: mainly: pair production and bremsstrahlung (supposed to be well known) 1 atmosphere = many steps most of the energy goes into e. m. cascade muon rate: related to Nstep down to Edecay (larger if primary is heavy)

Principle and aims of the Observatory • Large area on two sites (both hemispheres) 2 x 3000 km 2 → few tens of events/year/site above 1020 e. V ( if spectrum extrapolates in 1/Ea with a ~ 3 , i. e. no GZK cutoff) → no statistical ambiguity on the spectrum around 1020 → full sky coverage (point sources and extended structures) • Hybrid detection (ground array + fluorescence telescope) - better geometrical reconstruction (<1 deg) - cross-calibration of energy (sources of systematic errors are different !) • More possibilities for primary identification - traditional use of Xmax from fluorescence profile - structure of the front at ground → stage of evolution (again, possibilities of cross-checks) - window to “exotic” primaries (photon, neutrino)

The end…

Layout of the Southern Observatory Surface Array 3000 km 2 1600 water tanks (1. 5 km spacing) Fluorescence Detector 4 sites 6 Telescopes per site (30 x 30 deg 2)

Water Cherenkov tanks Communications antenna Electronics enclosure GPS antenna Solar panels Battery box 3 – nine inch photomultiplier tubes Plastic tank with 12 tons of water

Optical system corrector lens (aperture x 2) 440 PMT camera 1. 5° per pixel segmente d spherical mirror aperture box shutter filter UV pass safety curtain

Status of the Array (May 2005) ~ 800 tanks deployed (~ 730 sending data) stable running most of the time 2 telescopes in activity (10 % of the time) 1 more soon

Calibration (very simplified !) SD: use the Vertical Equivalent Muon - directly measured with hodoscope - indirectly from”muon hump” in the field (very large statistics !) - electron from muon decay within tank Also studied: dependence on atm. conditions, water level, etc… Problem: response to photons, electrons vertical muons all muons FD: various tools - lidar (probing along a laser beam) - infrared cloud monitor - central laser facility (seen from all tel. ) - ballons (atm. param. ) - drum calibration (uniform illumination) etc… Thick cloud Clear sky linear behavior LIDAR DATA Thin laye r

tra jec tor y 1

tra jec tor traj y 1 ect ory 2

tra jec tor traj y 1 ect ory 2 overlapping signals

Geometric improvement using hybrid detection Shower detector plane: well defined Position within SDP: problem if small angular range: t(c) = t 0 + Rp/c tan((c 0 - c)/2) 3 param. to be measured Solution(s) - stereo view by 2 telescopes (big showers only !) - one more constraint: time at ground (1 tank is enough !) better than 1 deg (direction) and 100 m (position) achievable in hybrid mode, even at low energy

An example of hybrid reconstruction SD points extrapolation Geometrical fit from FD only Hybrid fit

A big stereo hybrid event ! [Fick Plots]

Profile reconstruction with FD Cherenkov subtraction this event: intial viewing angle 15°, i. e. large direct Cherenkov contribution iterative procedure, converges in <4 steps; suggested energy here 2 Ee. V total observed Direct Ch. Scattered Ch. Gaisser-Hillas fit

Energy reconstruction with SD St at ist ica le rro rs on ly ! using the lateral profile to evaluate S(1000) to energy: under tuning (simulation+hybrid events)

The biggest SD event (up to now…) zenith angle 60 deg Energy of the order of 10 20 e. V (large uncertainty !)

First SD-FD energy comparison y r na i m i l e r P Not yet very precise, but clear correlation…

“young” and “old” showers as seen by SD “vertical” (13 deg) (long signal with multiple peaks) “horizontal” (76 deg) (muonic tail: very short peak) Evaluation of “age” + precise value of zenith angle (1 deg) indirect measurement of Xmax (less precise than FD, but more statistics !)

Sky coverage (galactic coord. )

Present rates Per day, in present configuration of SD (q < 60 deg, excluding edges and regions around “holes”) : ~500 events ~100 above 1018 e. V most of them fully reconstructible ~2 above 1019 e. V (but: preliminary energy scale !!! ) Hybrid: about 5 % of SD reconstructed events (regions not yet covered in front of telescopes) + many (FD & 1 or 2 SD stations) at low energy (with improved geometry from timing in SD station)

X=0 First interaction Modelling of shower development (2) 1 atmosphere depth First steps: big shower-to-shower fluctuations (model dependent !) X ~ 1000 g/cm 2 Electromagnetic cascade + muons (+ a few hadrons) Large Npart : quasi-deterministic evolution of densities (“universal” shape) Main effect of initial fluctuations: • Global translation of e. m. cascade • Modulation of muon rate

Different models… (from S. Ranchon’s thesis) First interaction : two extreme situations Many secondaries, with low energies One “leader” with a large xlab QGSJET gives more “nearly elastic” interactions Fraction of energy carried by the “leader” (most energetic secondary) Xmax is more delayed w. r. t. Xfirst

A possible parametrization of Xmax distribution Xmax-Xfirst int. AIRES simulation package hadronic model : QGSJET 01 protons, 1019 e. V Gaussian part : high multiplicity Exponential tail : contribution of “nearly elastic” processes Xmax Simulated protons at 2. 1018 e. V Fitting a convolution: gaussian * exponential Lexp = (1+e) LCR-air e: contribution of the tail of the Xmax-Xfirst distribution (0. 1 to 0. 5 ; model dependent !) (from S. Ranchon’s thesis) Remark: if large (e. g. proton), L is not too much sensitive to measurement error…

Dependence on the primary • Longitudinal profile : <Xmax> increases with Eprim (~ 50 g/cm 2 per decade) decreases with A (~ 80 g/cm 2 between p and Fe) Problem : modelling uncertainties on <Xmax> are comparable to this difference ! (and the bias may depend on energy…) Can the exponential tail give an useful information ? Two unknown functions of E : composition of CR and cross sections ! • Muon content : - Again : differences between models comparable with p – Fe difference (~ 30 % ? ) - difficult to measure (signal from muons mixed with electromagnetic contribution)

Conclusion • The components work well • Deployment in good shape (at least in summer) • Multiple tools for calibration • Large statistics; window at “low” energy • The hybrid concept is validated • still work to fully inter-calibrate • northern site to be built ! • modelling errors to be controlled: possible bias on energy; identification is difficult can we hope a stabilization of UHE model predictions ? first physics results this summer…