The KASCADE-Grande Experiment O. Sima 1 for the KASCADE-Grande Collaboration 2 1 University of Bucharest, Romania 2 https://web.ikp.kit.edu/kascade/ CSSP14 Sinaia 2014
Overview 1. KASCADE-Grande experimental facilities 2. Calibration and primary data analysis 3. Shower reconstruction 4. Standard energy and mass reconstruction 5. Energy spectrum and composition 6. S(500) and Constant Intensity Cut method 7. LOPES 8. KCDC 9. Conclusions
High energy cosmic rays Direct measurement: baloons, satellites Indirect measurements based on Extensive Air Showers (EAS) Konrad Bernlöhr
1. KASCADE-Grande experimental facilities Purpose study of cosmic rays in the energy range 10 16 10 18 ev - energy spectrum - mass composition Location Karlsruhe Institute of Technology (KIT Campus North) - 49.1 N, 8.4 E, 110 m a.s.l. 1022 g cm -2 - magnetic field: 47.8 mt, inclination 65 downward - multidetector system
KASCADE = KArlsruhe Shower Core and Array DEtector
dj A tr N j C pa(lg Ne, j,lg Nm, j E) lg( E) d lg( E) A Correlation between the truncated number of muons and the number of electrons in KASCADE (Astropart Phys 31 (2009) 86)
Light primaries spectrum All particles spectrum Main results of KASCADE: - energy spectrum in the range 10 14 10 17 ev - knee in the spectrum of light components (p, He, possibly CNO) at increasing energies - gradual increase of the heavy component with energy - no severe problem with data and analysis, consistent results - high energy model dependence still the limiting factor in the analysis of KASCADE data
KASCADE-Grande:measurements of air showers with energy 100 TeV - 1 EeV Grande Array KASCADE-Grande Grande station KASCADE Lopes Piccolo Array
Detector Particle Area[m 2 ] Threshold Grande Array (plastic scintillator) e/ + m 370 3 MeV Piccolo Array (plastic scintillator) e/ + m 80 3 MeV KASCADE Array (liquid scintillator) e/ 490 5 MeV (plastic scintillator) m 622 230 MeV Muon Tracking Detector (streamer tubes) m 128x4 800 MeV Calorimeter (liquid ionisation chambers) h 304x8 50 GeV Top Cluster (plastic scintillator) e/ 23 5 MeV Top Layer (liquid ionisation chambers) e/ + m 304 5 MeV Trigger Layer (plastic scintillator) m 208 490 MeV Multiwire Proportional Chambers m 129x2 2.4 GeV Limited Streamer Tubes m 250 2.4 GeV
37 detector stations (370 m 2 ) 18 trigger clusters (7 detectors)
2. Calibration and primary data analysis Calibration of Grande detectors The single particle spectrum compared to simulations Energy calibration in terms of the energy deposited by a vertically incident muon In shower analysis => Conversion of the energy deposited in the detector in particle density Lateral Distribution Function
In standard analysis Lateral Energy Correction Function LECF: LECF = function(r), r radial coordinate, independent of Standard LECF LDF for muons and electrons very important for shower reconstruction, for obtaining information on energy spectrum and mass composition
More realistic Lateral Energy Correction Function (LECF) LECF mean energy deposit per charged particle (e, m) - contributions to LECF from, p, n - radial dependence of LECF: fractional contribution of, p, n changes with r angular distribution of particle momentum => length of the trajectory in the detector - azimuthal dependence of LECF angular distribution of particle momentum => length of the trajectory in the detector S q Early O Late
-For the computation of LECF including all ingredients (particles, energies, angular distribution), efficient procedures to evaluate energy deposition are required - GEANT simulations - time consuming - not appropriate
Fast simulation procedure of energy deposition Method: Step 1 (data base) - full simulations with GEANT for e, m,, p, n for a set of - incident energies - incident angles - fit each GEANT distribution with a combination of simpler distribution functions (Landau, exponential, polynomial) - fit the dependence of the parameters of the resulting functions on the energy and the angle of incidence Step 2 (simulation) - find the parameters of the distribution functions corresponding to the particle, energy and angle of incidence -apply the Composition Method for Monte Carlo simulation of energy deposition by sampling from the simpler distributions Gain in computation speed: factor of 100 1000 in comparison with GEANT simulation Nucl. Instrum Methods A 638 (2011) 147
m e 1.2 GeV 50 GeV 50 TeV m e p n E(MeV) E(MeV) E(MeV) Electrons and muons q=0 o Electrons and muons q=40 o All particles, 50 GeV, q=40 o GEANT3.21 energy deposition (MeV)
GEANT distributions are decomposed into simpler distributions: - Landau - exponential - polynomial
Fit of the Landau peak Muons E=50 GeV 0, 20, 30, 40 o Angle dependence of the fit parameters
=> computation of LECF as a function of - incidence angle of the shower - radial distance from the shower core - azimuthal position Refinement: use of GEANT4, implement complete geometry of the hut and detector (A. Gherghel-Lascu) Correct density of particles in the observation level
Mapping of LDF in instrinsic shower plane Asymmetry of LDF in detector plane: shower development ( attenuation ) geometry effects geomagnetic field Information from the KASCADE-Grande detectors does not allow a complete integration of the lateral density function (LDF) around the shower core => deviations from azimuthal symmetry may be important - especially at large radial distances (S500) - muon density is reconstructed from KASCADE array (located in the NE corner of the Grande area) S Early O Late CSSP10 Sinaia 2010
CORSIKA simulation without the Earth s magnetic field => Projection onto the normal plane reduces the asymmetry, but does not eliminate it => The simple model not adequate Method 1: orthogonal projection Method 2: along particle momentum Method 3: triangulation based on arrival time Differences between projection methods Still important asymmetry
Correction Factor: Exponential attenuation coefficient -function of angle -function of radial coordinate Symmetry restored
Test of the LDF Reconstruction: Comparison with density reconstructed by KASCADE
Nucl. Instrum. Methods A 620 (2010) 202 Energy calibration: uncertainty 15-17%
Arrival Time and Shower Direction Timing uncertainty: 2 ns Direction: 0.8-1.2 degree Core position: 5-8 m Shower size Shower core position Grande versus KASCADE Arrival direction Grande versus KASCADE Nucl. Instrum. Methods A 620 (2010) 202
3. Shower reconstruction - Direction - Energy of the primary particle - Mass of the primary particle - Primary experimental information - KASCADE-Grande: energy deposition in detectors => charged particle density - Arrival time - KASCADE: muon density, electron density - High energy muons map of the atmospheric depth of production - Radio signal (LOPES)
The recorded data is used to infer information on the primary particle QGSJETII Detector calibration Conversion of detector signal in particle density Mapp the density in intrinsic shower plane Construct observables Inference of the parameters of the primary particle Crosschecks between results are intended/possible
The recorded data is used to infer information on the primary particle QGSJETII Detector calibration Conversion of detector signal in particle density Mapp the density in intrinsic shower plane Construct observables Inference of the parameters of the primary particle Crosschecks between results are intended/possible
Standard shower reconstruction: KRETA calibrated detector signal converted to charged particle density using LECF (mean energy deposit per charged particle) muon density obtained by fitting the data from KASCADE muon detectors electron density obtained by fitting charged particle density and muon density extrapolated from KASCADE to Grande detectors using specific LDFs primary energy evaluated using the muon and the electron size of the shower primary particle identification based on the correlation of the electron size and the muon size of the shower LECF, LDF, energy estimator, mass estimator obtained from the analysis of simulated showers using CORSIKA
Lateral Density Function (LDF) Energy deposition in detectors converted in particle density using LECF Particle density in intrinsic shower core information about primary energy and about primary composition NKG Linsley Lagutin Polynomial
S 100 S 600
Fiducial Area Trigger Efficiency Reconstruction Efficiency
4. Standard energy and mass reconstruction - Number of charged particles: power law in function of energy - Logarithm of E linear function of logarithm of N ch, but: - Dependence on primary mass Log 10 (E)=a(m) log 10 (N ch ) + b(m) - a, b dependence on primary mass through the ratio R=(N ch /N m ): a=a p +(a Fe -a p ) k; b=b p +(b Fe -b p ) k k=(log 10 (R)-log 10 (R p )/(log 10 (R Fe )-log 10 (R p ) R p =(N ch /N m ) p, R Fe =(N ch /N m ) Fe - The coefficients depend on the zenith angle (muons and electrons attenuate differently) Coefficients determined by simulations Astropart. Phys. 36 (2012) 183 Tests: application to simulated showers
Reconstructed vs simulated number of showers in energy bins Epos simulations QGSJet-II used used for calibration Astropart. Phys. 36 (2012) 183
Reconstructed vs simulated energy average and standard deviation Astropart. Phys. 36 (2012) 183
5. Energy spectrum and composition All particle spectrum, calibration based on QGSJet-II - With Epos 1.99 calibration, spectrum goes down by 10% Astropart. Phys. 36 (2012) 183
Conclusions from Astropart. Phys. 36 (2012) 183: Measured all particle spectrum with a systematic uncertainty <15% Spectrum between knee and 10 18 ev: - Not a single power law - Spectrum hardening above 10 16 ev - Spectrum steepening around 10 17 ev Reconstruction of energy spectra of groups of primary mass: - Mass separation k coefficient k=(log 10 (R)-log 10 (R p )/(log 10 (R Fe )-log 10 (R p ) R p =(N ch /N m ) p, R Fe =(N ch /N m ) Fe => Proton primary: k=0 Iron primary: k=1 Electron rich showers: light primary Electron poor showers: heavy primary
Reconstructed spectrum for electron poor [k>(k Si +k C )/2] and electron rich showers Phys. Rev. Lett. 107 (2011) 171104
Steepening of the all particle spectrum around 10 17 ev First evidence that this knee-like break in the spectrum is due to the heavy primaries Spectrum of light primaries re-evaluated Phys. Rev. D 87 (2013) 081101 Definition of light primaries: electron rich showers, with small k, k < (k C +k He )/2 - Improved statistics: - Additional observation time - Increased fiducial area
=> Ankle like structure in the spectrum of light primaries at about 10 17 ev (energy above the knee of the heavy primaries)
Spectrum of the light component of cosmic rays - Slope change from -3.25 to -2.79 at around 10 17 ev - Possible indication of the transition from galactic to extragalactic cosmic rays
Dependence of reconstructed spectrum on interaction models (low energy interaction model: Fluka) Bands in the figure denote the effect of systematic uncertainties Adv. Space Research 53 (2014) 1456
Model dependence: EPOS produces larger number of muons than QGSJet or SIBYLL, with SIBYLL correponding to the lowest number of electrons and muons at sea level => EPOS gives lower values of energy than QGSJet, SIBYLL higher Reconstruction of the flux of showers simulated with SIBYLL, QGSJet and EPOS using calibration based on QGSJet Mixed composition with 5 primaries with equal contribution Adv. Space Research 53 (2014) 1456
Differences between the composition based on SIBYLL and EPOS However, the spectrum behaviour for the electron rich and electron poor showers is the same => Knee around 10 17 ev for the electron poor (heavy) primary => Ankle for the electron rich (light) primary
Slight sensitivity to the k value for the selection of electron poor vs electron rich samples Adv. Space Research 53 (2014) 1456
Nucl. Phys. B Proc. Supplement (2014) in press
Nucl. Phys. B Proc. Supplement (2014) in press
Nucl. Phys. B Proc. Supplement (2014) in press
Conclusions concerning the spectrum: - KASCADE: knee of the light primaries at about 3 10 15 ev - Spectrum hardening at 10 16 ev due to medium mass primaries - KASCADE-Grande: knee of heavy primary around 10 17 ev - Ankle due to the light primaries at 1-2 10 17 ev - Mixed composition between 10 15 8 10 17 ev
6. S(500) and Constant Intensity Cut method Independent approach G. Toma, ICRC 2013 - For KASCADE-Grande array, at 500 m from the core the density of charged particles S(500) is independent of primary mass S(500) energy estimator Calibration by Monte Carlo simulation
Good accord of the reconstructed energy by S(500) and by the standard method for Corsika simulated showers Discrepancy in the case of experimental data Inconsistency between simulations and data higher muon multiplicity?
7. LOPES Radio emission from air showers - Geomagnetic deflection of electrons and Positrons thickness of the shower front 1 m, frequencies of tens of MHz - Askaryan effect, charge excess Advantages: - High duty cycle - Precision - E: 10 17-10 18 ev Nature 435 (2005) 03614 - Correlation of radio emission with EAS Detected by KASCADE-Grande
30 dipole antennas
F. Schroeder, ECRS 2012 Energy reconstruction
Radio detection of EAS implemented in other high energy cosmic ray installations
8.KCDC KASCADE-Grande what s next? KASCADE and KASCADE-Grande closed and dismantled
KASCADE Cosmic Data Center KCDC
9. Conclusions KASCADE-Grande a unique high energy cosmic ray detector - multidetector, hadron, electromagnetic, muon, radio components - Provided energy spectrum and mass composition of high energy cosmic rays with a well controlled uncertainty - Observed the spectrum structures heavy ion knee, ankle of the light component, besides the knee due to the light primaries at 3 10 15 ev - The results do not require new interaction features at the energies up to 10 18 ev - Implemented and tested novel experimental techniques, refined analysis procedures - Provided information for testing the models of acceleration and propagation of cosmic rays
Thank you!
p, 3.16 10 17 ev, q=45 o from North CSSP10 Sinaia 2010
Fe, 5.62 10 17 ev, q=45 o from North CSSP10 Sinaia 2010
CSSP10 Sinaia 2010