Extensive Air Shower Observation for Energy < and Related Topics

Transcription

1 Frontiers of Cosmic Ray Science 277 Extensive Air Shower Observation for Energy < 17 and Related Topics ev Masato Takita Institute for Cosmic Ray Research, the University of Tokyo, Kashiwa, Chiba , Japan Abstract The post conference proceedings of this rapporteur talk will cover the sessions HE 1.1 (Extensive Air Shower Observation for Energy < 17 ev), HE 1.2 (Theory and Simulation in the Atmosphere), HE 1.5 (Instrumentation and New Projects, excepting UHECR) and HE 3.1 (Hadron Interactions). All of the contributed papers and talks in these sessions included new and interesting results that deserve discussions and presentations in rapporteur talks. It is simply my own interest and limitations that the scope of this rapporteur proceedings paper is narrowed. 1. Introduction To start with statistics, there are 58 contributed papers in HE 1.1, 23 papers in HE 1.2, 93 papers in HE 1.5, and 20 papers in HE 3.1, demonstrating that these particular sessions are very active. These papers cover a wide range of important topics, containing cosmic-ray energy spectrum and composition measurement obtained with air shower experiments, testing and development of hadronic interaction models, experimental research on particle nuclear interactions by accelerator results at several GeV/nucleon as well as by the results from emulsion chambers at mountain altitude, exotic events and new particle searches in different types of cosmic ray data, new experimental technique and development and future projects. All of these contributed papers and talks included new and interesting results that deserve discussions and presentations in rapporteur talks. It is simply my own interest and limitations that the scope of this rapporteur proceedings paper is narrowed. It should be noted that this summary does not include TeV gamma-ray astronomy, direct observations of high-energy cosmic rays, nor ultra high-energy cosmic rays, nor the origin of the primary cosmic rays, nor their propagation to the Earth. pp c 2004 by Universal Academy Press, Inc.

2 Extensive Air Shower Observation for Energy < 17 ev The cosmic-ray energy spectrum obeys a power-law spectrum with 3 breaks. The observed spectral breaks - the knee at several 15 ev where the differential power-law index changes from -2.7 to -3.0 and the ankle at several 18 ev beyond which the energy spectrum becomes flatter extending up to 20 ev, and the Greisen-Zatsepin-Kuzmin cut off around 20 ev - are still controversial. The knee was first observed by the MSU group in the electromagnetic component more than 40 years ago [31]. Charged cosmic rays < 17 ev detected at the Earth provide no information on their initial directions, due to their diffusion in the interstellar medium of the Galaxy filled with irregular magnetic fields of the order of µgauss. The energy spectrum is affected by their propagation in the Galaxy. The most natural explanation of their origin is the the energetic objects in the Galaxy, such as supernova remnants. The Fermi shock acceleration might be the most plausible acceleration mechanism. The chemical composition and energy spectrum of cosmic rays provides us with very important clues on their origin, however, direct measurements in satellite- or balloon-borne experiments are available only up to 0 TeV, because their payload is limited and the cosmic-ray event rate drops rapidly as their energies go up. Therefore, we have to measure the chemical composition and energy spectrum of cosmic rays above 0 TeV by means of ground-based experiments, such as extensive air shower arrays. The goal of the experiments is to measure the elemental composition as a function of energy in the knee region, or the energy spectra for separate mass groups. It is particularly important to settle down if the break energy is proportional to atomic number Z or proportional to mass number A All-particle energy spectrum Figure 1 shows a compilation of the all-particle energy spectrum around the knee region. In the knee region, the air shower size reaches its maximum at altitude higher than 4000 m above sea level, therefore, the all-particle energy spectrum measured at highland solely becomes free from dependence on the chemical composition in the primary cosmic rays. The energy spectrum dn/de is multiplied by E 2.5 to magnify the deviation from a simple power-law spectrum. This makes the spectral shape resemble a human knee, leading to the anatomical name knee. At first sight, the absolute flux values measured by the various experiments might look inconsistent among them. However, they are consistent with one another taking it into account that the possible energy scale uncertainties among the experiments amount presumably

3 all particles dj/de x E [ev m sr s ] Tibet III(2003) 16 KASCADE(2003) EAS TOP BASJE MAS(2003) Akeno BASJE CASA MIA 15 DICE HEGRA JACEE MSU proton RUNJOB SOKOL Sphere TACT Tunka Tibet I 14 Primary Energy [ev] Fig. 1. All-particle energy spectrum around the knee region. References are Tibet-III(2003);[7], KASCADE(2003);[38],EAS-TOP;[1], BASJE-MAS(2003);[35], Akeno;[34], CASA-MIA;[20], DICE;[42], HEGRA;[12], JACEE;[13], MSU;[17], proton;[22], RUNJOB;[], SOKOL;[28], Sphere;[8], TACT;[9], Tunka;[21], Tibet-I;[4]. to ±20% or more, and that all the results do not employ the same simulation code, interaction model and primary composition model. The general consensus is that the all the experiments suggest the existence of a sharp spectral break around several 15 ev. There is now convincing evidence that the chemical composition becomes heavier through the knee region, as < lna > increases by roughly 2 in the knee region as shown in Fig. 2. At the conference, impressive results among others are presented from the KASCADE experiment and the Tibet ASgamma experiment, providing the energy spectra for separate mass numbers in the knee region KASCADE Results The KASCADE air shower array [33],[38] is a combination of detectors which are capable of measuring the electromagnetic, muonic, and hadronic components of individual air showers. They measure the the energy spectrum of unaccompanied hadrons with the

4 280 Fig. 2. < lna > as a function of energy from various experiments. Results from SPASE-2/AMANDA-B [36] and EAS-TOP [44] are added tothe original BASJE compilation [43]. large hadron calorimeter of the KASCADE experiment. Unaccompanied hadrons are cosmic-ray induced events for which solely one hadron has been recorded at ground level. They offer an opportunity to study detailed features of hadronic interactions in the atmosphere and the primary proton spectrum in the energy range from 0 GeV to 5 15 ev. The definition of an unaccompanied hadron in their analysis is: Only one hadron with energy of at least 50 GeV with zenith angle smaller than 30 degrees. Proton separation capability is demonstrated in Fig. 3. The Monte Carlo simulation is carried out using CORSIKA with the GHEISHA and QGSJET01 hadronic interaction codes. The unaccompanied hadrons interact only a few times in the 11 interaction length thick atmosphere before they reach the detector. Integrated over all relevant energies, the average number of interactions is 3.6±1.9 for proton and 6.6±2.3 for iron induced air showers. Both numbers are smaller than the corresponding ones for all the air showers, 6.4±1.8 for proton and 7.6±2.2 for iron, respectively. Between October 1996 and October 2001, they accumulated more than 3 8 with a least one reconstructed hadron in the hadron calorimeter. The measured unaccompanied hadron flux is also shown in Fig. 3. Based on the simulations the unaccompanied hadron flux is converted to the proton energy spectrum at the top of the atmosphere, as shown in Fig. 4. There is no indication for steepening in the proton spectrum between TeV and 0 TeV. The derived proton energy spectrum agrees well with direct and indirect measurements. In the knee region, the statistical error bars are unfortunately too large to give any meaningful result.

5 281 Fig. 3. KASCADE: [33]. Left: Number of interactions versus hadron energy at ground level. Right: The measured unaccompanied hadron flux at ground level. Fig. 4. KASCADE [33]. Left: Number of unaccompanied hadrons versus primary energy for different elemental groups. Right: Primary proton flux reconstructed from unaccompanied hadrons compared to results of direct measurements. The line represents a fit tothe measurements. On the other hand, the KASCADE experiment measures the electronic and muonic components of air showers in the knee region with high statistics. On the basis of these measured data, corresponding two-dimensional Ne vs. Nµ shower size spectra are obtained, as shown in Fig. 5. The main point is, that they unfold the two-dimensional size spectrum (muon and electron number) of the measured air showers. The correlation of the muon and electron number gives an additional constraint to infer individual spectra leading to a unique solution for a given set of mass groups. The measured energy spectra are shown in Fig. 6. They have investigated possible systematics by using two different hadronic interaction codes underlying the unfolding (for the estimate of the kernel-function

6 282 Fig. 5. KASCADE two-dimensional Ne vs. Nµ shower size spectra [38]. The zenith angle range of the shower is from 0 to 18 degrees. in the Fredholm equations), by using different unfolding methods (Bayesian unfolding, unfolding by a Gold-algorithm, etc.) as shown in Fig. 7 (left), and by using different sets of the measured data. They found the following statements keep very stable. The flux and slope of the all-particle spectrum is stably measured (which is just the sum of the five resulting spectra for the various mass groups). There is a kink of the light particles between 2 and 6 PeV. There is no distinct kink of heavy primaries between 1 and PeV. The position of the kink of proton and helium differ by about factor two rather than by a factor of 4 as shown in Fig. 7 (right). But they also found that the relative abundances of the various primaries depend very much on the chosen interaction model not less than 15%. And they found that this systematic uncertainty is mainly based on the fact, that the interaction models cannot describe the correlation of the muon to electron number on event-by-event basis; plain muon or electron size spectra alone can be reproduced by the simulations if a certain composition is assumed. Detailed systematic errors are not included at the figures presented at the ICRC2003, as they are still under investigations. Using the QGSJET model, the results of the unfolding procedure are confirmed by the analyses of the unaccompanied hadrons (using also QGSJET).

7 283 Fig. 6. KASCADE [38] results of energy spectra by the Gold unfolding procedure. The given error bars reflect the statistical errors due to the measurement and simulation. The all-particle spectrum as well as the spectra for light elements are displayed. Systematic errors for the all-particle spectrum due to the applied method are indicated by the shaded area Tibet ASgamma Results A hybrid experiment consisting of the emulsion chambers (ECs) with burst detectors (BDs) and the air shower array (AS array) of 298 scintillators with 15 m spacing (36,900 m 2 ) was carried out at Tibet Yangbajing (4,300 m a.s.l.) from 1996 through The total area of ECs was 80m 2, having a 50 cm 40 cm area and 14 c.u. thickness (lead) in each unit, where X-ray films were inserted at 4, 6, 8,, 12 and 14 c.u. depths. The BDs with the same area were placed just below ECs. Each BD contains a plastic scintillator with the size of 160 cm 50 cm 2 cm, namely, 4 ECs were placed above one burst detector. Thus, 400 blocks of ECs and 0 BDs in total were used in this experiment. The X-ray films in ECs were exchanged by new ones every year to suppress the background. The AS array is triggered by BDs. For each γ-family event detected in ECs, its accompanied air shower can be found using information; (1) the location of the burst, (2) the time matching of the burst with the air shower and (3) the comparison of arrival direction of the gamma family with that of the air shower. The details of the method of analysis are based on Monte Carlo simulations (Cosmos code) with two different primary composition models and neural network analysis (ANN) to identify the proton induced events. The statistics is increased by about factor 3 at ICRC2003 compared with the last one at ICRC2001

8 284 Fig. 7. KASCADE [38]: Comparison of the deconvoluted H and He spectra using Bayesian and Gold unfolding, with solid lines marking the systematic uncertainty of the Gold algorithm (left panel). Individual spectra as a function of the rigidity E/Z. The knee position of the H and He spectra are nearly the same position (right panel). by completing the EC analysis. The ANN target value distribution using 5 event characteristics, N γ (multiplicity of a family), ΣE γ (energy sum of a family), <R γ >(mean lateral spread of a family), <ER γ >(mean lateral spread of the family energy) and N e (shower size) is shown in Fig. 8 and Fig. 9 for the assumptions of heavy dominant composition (HD) and proton dominant one (PD), respectively, using the data of whole energy range, where the training was made with t=0 for proton primary events and t=1 for others. The arrow with symbol T in the figure indicates the cut off value for the separation of proton-like and heavy-like events defined so as to cancel the contaminating numbers of events in both regions each other. The energy dependence of the value T was also examined using simulated data to reduce the error involved in the reduction of the contamination. They present a result from 178 family events with accompanied air showers, among which 112 and 66 events are assigned as proton-like and heavy-like events, respectively, in case of the HD model, while 130 and 48 events in case of the PD model. The proton spectrum obtained in the analysis is shown in Fig., where plots labeled as HD are obtained from the analysis based on heavy dominant composition model, while those labeled as PD are based on proton dominant one. These two results agree well within statistical errors showing there is no significant

9 285 Fig. 8. [6]. ANN target value distribution (Heavy Dominant model) for Tibet ASgamma primary composition model dependence in their analysis. A fit to the low-energy proton spectrum data is made by T.K.Gaisser [18] using recent magnetic spectrometer measurements leading to a power index in high energy region, which seems to accommodate direct observations up to a few hundred TeV, however, they have observed steepening of the proton spectrum in the knee region. The power index in the energy region > 15 ev is estimated by assuming the HD primary composition to be -3.14±0. deviating by 3.9 σ from the extrapolation of the Gaisser s fit. Another estimate based on the PD model also results in power index of -3.06±0.09 being 3.6σ deviation. If they estimate the break point of the proton spectrum around 200 TeV from Fig. and assume other components also have the break points proportional to their atomic number, then the knee of the all-particle spectrum can be interpreted as the break point of the iron component Comparison between KASCADE and Tibet ASgamma results Figure 11 shows the comparison between the Tibet ASgamma and KAS- CADE proton fluxes. To reduce the energy scale uncertainties, the proton/allparticle flux ratio is shown in Fig. 12. The KASCADE unfolding results are presently starting at 1 PeV, the statistics of Tibet ASgamma at the proton selection above 5 PeV is still small, and between 0 TeV and 500 TeV the direct measurements are very rare and with large statistical and systematic uncertainties. However, it may be difficult

10 286 Fig. 9. [6]. ANN target value distribution (Proton Dominant model) for Tibet ASgamma to explain the possible discrepancy between the Tibet ASgamma and KASCADE proton fluxes between 1 PeV and 5 PeV, if one considers only the statistical errors. Extensive study of possible systematic errors are under way by both groups. Another interpretation is the theory: Nobody forbids us to think about two kinks in the proton spectrum, as it can occur regarding different sources, e.g. of supernovas of different initial masses (see S. Sveshnikova [41]). Figure 13 summarizes the helium energy spectra obtained by various experiments. The KASCADE helium flux [38] appears to be away from the smooth extrapolation of the direct measurements, indicating helium-rich energy spectrum, when one takes into account the statistical errors only. The published Tibet ASgamma helium spectrum [5] is smoothly connected with the direct measurements. Again, possible systematic errors in the KASCADE measurement should be estimated in order to draw any conclusions. 3. Hadron Interactions Let us first mention the relatively low-energy hadron interaction issues, mainly relevant to atmospheric neutrino flux calculations. The fluxes of atmospheric neutrinos have been of great interest, as they are sensitive to neutrino oscillations. The neutrino oscillations are predicted, starting from the known fluxes of primary cosmic rays at the top of the atmosphere. Hadronic interactions are then simulated by Monte Carlo methods as these particles strike the atmo-

11 287 Fig.. Proton spectrum obtained by Tibet hybrid experiment [6]. Errors are statistical only. sphere and the cascade air showers are traced till the particles produce neutrinos after their decays. Solar activities and geomagnetic effects are also taken into account. This technique relies on a knowledge of the production of hadrons in collisions with air molecules, and this element of the simulation dominates the uncertainty in the cosmic-ray fluxes. Such basic data of hadronic productions have been collected by the NA49 [14] and the HARP [37] experiments at CERN. The layout of the NA49 main detectors is shown in Fig. 14. The NA49 experiment uses large time projection chambers to measure the momenta and identify the type of particles produced in high energy collisions. The data with the carbon target (the closest convenient isoscalar nucleus to the nitrogen and oxygen in the air) were collected in a week run during The target is cylindrical, 6 mm in diameter and 6mm long. Data are collected exclusively with the differential Cherenkov set to single out protons. Two separate beam momenta are measured. The first set with beam momentum 158 GeV/c contains 500,000 triggers and the second set with beam momentum 0 GeV/c contains 160,000 events. Approximately half of the triggered events are induced by real physical interactions. These will be analyzed shortly, and are intended to be used to make the hadron production models used in atmospheric neutrino flux calculations more precise. The HARP experiment is located at CERN and uses a secondary beam generated by CERN PS. The experiment was constructed during 2000 and the beginning of 2001 and started physics data-taking during the summer of 2001

12 288 Proton dj/de x E 2.5 [ev 1.5 m 2 s 1 sr 1 ] Tibet (2003) KASCADE(2003) unaccompanied hadron EAS TOP BASJE JACEE Kawamura RUNJOB SOKOL USSR Primary Energy [ev] Fig. 11. Comparison between the KASCADE and Tibet ASgamma proton spectrum. and then continued throughout Figure 15 shows a schematic view of the HARP detector layout. The target is located inside the time projection chamber (TPC) allowing high transverse momentum and even backward-going particles to be measured. The TPC provides tracking, momentum measurement (0.7 Tesla solenoidal magnetic field), and de/dx particle identification. Particle identification is supplemented, in this large angle region, by a time-of-flight (TOF) system, based on resistive plate chambers. Data are collected for a set of beam momenta (including 3, 5, 8,, 12, 15 GeV/c) and for a variety of targets ranging from beryllium to lead. Data are also taken during 2002 with liquid cryogenic targets: hydrogen, deuterium, nitrogen and oxygen. Usually the hadron production data used for determination of the atmospheric neutrino flux employs an isoscalar target similar to nitrogen but more convenient to handle (for example, carbon or beryllium). With the HARP data it will be possible to input data taken with a nitrogen target directly. The analysis of this data is on-going. Now, we will turn to the high-energy hadron interaction issues. The Relativistic Heavy Ion Collider (RHIC )at Brookhaven accelerates protons or nuclei up to 250 Z/A GeV in each beam, corresponding to a cosmic-ray proton with 133 TeV collision with a proton at rest. The Tevatron Collider at the Fermilab at 0.95 TeV beam energy produces proton-antiproton collisions at energies equivalent to a 2 15 ev cosmic-ray proton. The Large Hadron Collider (LHC) at CERN which is planned to start operation in 2007, will be a collider of a 7 TeV proton with a 7 TeV proton, equivalent to cosmic-ray proton over 17 ev. The LHC program includes nucleus-nucleus collisions, too. It is obvious that the LHC

13 Tibet KASCADE 0.2 (dj/de)p (dj/de)all / Primary Energy[eV] 17 Fig. 12. Comparison between the KASCADE and Tibet ASgamma proton/all-particle flux ratio. data will be valuable for the cosmic-ray Monte Carlo hadron interaction simulations. Until the LHC data is available in 4 years, the cosmic-ray community will have to work with the relevant data which will be produced at RHIC and Tevatron Collider. On the other hand, the air shower experiments will continue investigating the knee energy region problem, based on the best available knowledge on the interaction models and primary comic-ray composition models. As an example of exotic phenomena, the Centauro-I event should be mentioned. There appears now a possible new interpretation from members of the group [30] on the Centauro-I event [29] detected by the Chacaltaya experiment, which continues to be very unlikely fluctuations or a new type of particles/ interactions compared with hadron gamma families with large statistics now. The new interpretation is quoted from the reference [30]: Hence, if the detector had been ideal, then the most plausible explanation for the origin of Centauro-I, would have been SQM (strange quark matter). In the real situation of the experiment, the most likely cause of the event I-12 observed in the chamber CH-15 would be the Chacaltaya detector problem. An agreed-upon interpretation is awaited. 4. Simulations The CORSIKA simulation code with QGSJET hadronic interaction models has become a de-facto-standard model for air shower simulations on which experimental data analyses are based. This is an important progress that allows

14 290 He dj/de x E 2.5 [ev 1.5 m 2 s 1 sr 1 ] Tibet BD(Corsika+HD) Tibet BD(Corsika+PD) KASCADE(2003) EAS TOP(2003) BASJE Ivanenko JACEE Kawamura RUNJOB Primary Energy [ev] Fig. 13. Comparison between the KASCADE and Tibet ASgamma helium flux. to compare different experiments. There is possible danger lurking, however, as one relies too much on a single code and model. We should encourage alternative simulation codes, for example, COSMOS, to be developed and compared with the CORSIKA code for the healthy development of air shower simulation codes. Continuous efforts to understand possible systematic uncertainties from the improper modeling will be important. One of the conspicuous presentations among others is the discussion [25] on the influence of low-energy (Elab < 0 GeV) hadronic interaction models employed in the CORSIKA code. In the past, mostly GHEISHA routines have been used in air shower simulations with CORSIKA. However, it is known that GEANT-GHEISHA suffers from incompleteness in handling the reaction kinematics properly. For example, in air shower simulations using GHEISHA, the sum of the energy of the secondary particles and the energy deposit is often larger than the primary energy by several %, depending on the primary energy and the lowenergy threshold (typically 300 MeV) above which hadronic particles are followed. As an alternative to GHEISHA, the hadronic event generator of the FLUKA 2002 code has been coupled with CORSIKA. Independently, correction patches for GHEISHA became available which improve energy and momentum consevation, but do not change basic properties like particle multiplicities or differential cross sections. For p- 9 Be interactions, several experimental data sets are available at Elab 20 GeV. Figure 16 shows the distributions of secondary mesons as a function of xlab= ptot/pbeam. Generally, the experimental data are well reproduced by FLUKA, while GHEISHA produces significantly less mesons at xlab 0.15

15 291 Fig. 14. Schematic layout of the main components of the NA49 experiment used in the hadroproduction measurement [14]. and slightly more in the region of xlab This feature holds also for other types of hadronic collisions with 14 N targets. As GHEISHA and FLUKA predict different momentum distributions of secondary pions, the spectra of muons with Elab < 30 GeV depend on the employed low-energy model. Figure 17 shows muon energy spectra for several combinations of low- and high-energy models with transition energies of 80 GeV and 1.5 TeV. The largest differences between the energy spectra amount to 15% at Eµ 0.8 GeV correlated to the differences in the predicted distributions of pions around xlab=0.15. Its influence on the hadronic and muonic component is obvious, while the electron densities in the simulation show no significant dependence of the employed low energy model. For CORSIKA applications that are sensitive to low-energy muon numbers and energy spectra, the replacement of GHEISHA by FLUKA is recommended. 5. Future Projects There are many proposals and new projects excepting the UHECR energy region presented at the conference. Several of them will be shortly introduced The ARGO-YBJ experiment The ARGO-YBJ experiment [2], [15], [24], [27], [32], [40] currently under construction at the Yangbajing Laboratory located 4300 m a.s.l. in Tibet, China, as an Italian-Chinese collaboration project. The detector is composed of a single-

16 292 Fig. 15. Schematic of the components of the HARP detector [37]. layer of 1848 Resistive Plate Chambers (RPCs) operated in streamer mode for a total instrumented area of 6500 m 2. A lead converter 0.5 cm thick will cover uniformly the RPC plane to increase the number of charged particles by conversion of shower photons and to reduce the time spread of the shower front. The experiment aims at detecting air showers produced by primary cosmic rays with energy ranging from 0 GeV to TeV. The gas mixture used in the experiment, C 2 H 2 F 4 / i-c 4 H /Ar/SF 6 = 54.7 / 30 / 15 / 0.3, requires an operating voltage of kv. Presently, 36 clusters of the central detector carpet, corresponding to 1600 m 2 have been installed and partially put into operation for debugging and certification. It will start full-scale data-taking in The ASHRA experiment ASHRA (All-sky Survey High Resolution Air-shower telescope) [3], [11] can take an image of an air shower through two kinds of yielded lights, Cherenkov and fluorescence, with 1 arcmin angular resolution in the all sky coverage of field of view. These features of ASHRA can provide the systematic exploration into TeV gamma rays and precise measurement for arrival directions of EHECRs. The ASHRA observational station consists of 12 light collection telescopes covering all sky (50 degree 50 degrees) with totally 80 mega pixels in the CMOS sensor arrays with IIT. The station site candidates are currently locations near the summits of the three mountains of Mauna Loa, Hualalai, and Mauna Kea on the Hawaii Big Island. In the first step, they are planning to install one full station including 12 telescopes at the site near the Mauna Loa summit and 4 telescopes

17 293 Fig. 16. Distribution of secondary particle momenta xlab = ptot/pbeam in p- 9 Be collisions at plab = 24 GeV/c [25]. Left: Pions. Right: Kaons. The experimental data points are derived from the measurements. in another station on the top of Hualalai which is distant from Mauna Loa by 35 km (the ASHRA-1). The ASHRA-1 is scheduled to start data-taking in In the second step, they plan to enhance the Hualalai site into one full station and installing the third station at the site on the higher side of Mauna Kea (ASHRA- 2) The IceTop experiment The IceTop experiment [19] is a km 2 air shower array consisting of 80 pairs of ice Cherenkov tanks as an integral part of the design of the IceCube neutrino telescope at the South Pole. The solid surface above a neutrino telescope in deep ice makes possible the construction of a surface array for veto and calibration purposes. The resulting three-dimensional array is available for studies of cosmic rays up to the EeV range. The surface component of IceCube will consist of two tanks near the top of each hole, forming a km 2 array of 80 stations on a 125 m triangular grid. Each tank will be 2 m in diameter filled with ice to a depth of one meter. Each tank will be instrumented with two digital optical modules, one operating at low gain and the other at high gain to achieve dynamic range > 5, with some overlap for calibration. Combined analysis by IceCube (muon component) and IceTop (electromagnetic component) will allows to study the chemical composition of primary cosmic rays up to EeV.

18 294 Fig. 17. Energy spectra of muons arriving at detecter level (1 m a.s.l.) for primary protons of 14 and 15 ev, vertical incidence [25]. Left: QGSJET 01 combined with different low-energy models and transition energies. Right: QGSJET 01 and SIBYLL 2.1 combined with FLUKA at transition energies of 80 GeV and 1.5 TeV The KASCADE-Grande experiment A scintillator array (Grande) of large collecting area (700 m 700 m) has been set up at Forschungszentrum Karlsruhe in Germany to operate jointly with the existing KASCADE experiment. The enlarged air shower experiment provides comprehensive observations of cosmic rays in energy range from 0.1 PeV to 1 EeV., i.e., a full coverage of the primary energy region around the knee region. The major goal of KASCADE-Grande [16],[23] is the observation of the iron knee in the cosmic-ray energy spectrum around 0 PeV, which is expected following recent KASCADE observations where positions of the knees of individual mass groups suggest a rigidity dependence [38]. Additionally, the validity of hadronic interaction models employed in CORSIKA Monte Carlo simulations of ultra-high energy air showers will be tested with KASCADE-Grande. Investigations of radio emission in high-energy air showers will be made with a further upgrade of the experimental set-up by installing an array of broad-band antennas provided by the LOPES collaboration [26]. The existing multi-detector experiment KASCADE (200 m 200 m, 252 stations in total), which takes data since 1996, was recently extended to KASCADE-Grande by installing a large array of 37 stations consisting of m 2 scintillation detectors, with an average spacing of 137 m. The scintillators were taken from the former EAS-TOP experiment. The stations comprise 16 photomultiplier tubes each providing a wide dynamic range from 1/3 to 30,000 charged particles per station for the reconstruction of particle densities and timing measurements. The signals are amplified and shaped inside the Grande stations, and after transmission to a central DAQ station digitized in peak sensitive ADCs. KASCADE-Grande has a sensitive area of more than 0.5 km 2 and operates jointly with the existing KASCADE. The

19 joint measurements are ensured by an additional cluster (Piccolo) close to the center of KASCADE-Grande for trigger purposes. Piccolo consists of 8 m 2 stations equipped with plastic scintillators. A core position resolution of 15 m and an angular resolution of 0.5 degrees will be reached by the present set-up. A common fit to the energy deposits with the relative muon to electron ratio as additional free parameter enables a resolution of electron and muon numbers in the order of 15% and 25%, respectively, for primary energies of 0 PeV. 6. Summary The KASCADE group gave a preliminary result on the cosmic-ray proton, helium, CNO and all-particle energy spectra in the knee region, where the position of the kink of proton and helium differ by about factor two rather than by a factor of 4. The KASCADE result also suggests that the knee in the all-particle energy spectrum should be caused by the light nucleus and that there should be no distinct kink of heavy primaries between 1 and PeV. The KASCADE-Grande started operation to observe the iron knee around 0 PeV. Meanwhile, the Tibet ASgama experiment also gave a preliminary result on the proton spectrum in the knee region measured by the hybrid method, together with the all-particle spectrum. A fit to the low-energy proton spectrum is made by T.K.Gaisser using recent magnetic spectrometer measurements leading to a power index in high energy region, which seems to accommodate direct observations up to a few hundred TeV, however, Tibet ASgamma observed steepening of the proton spectrum in the knee region. The power index in the energy region > 15 ev is estimated by assuming a heavy-dominant primary composition model to be -3.14±0. deviating by 3.9 σ from the extrapolation of the Gaisser s fit. Another estimate based on a proton-dominant primary composition model also results in power index of -3.06±0.09 being 3.6σ deviation from the Gaisser s fit. If they estimate the break point of the proton spectrum around 200 TeV and assume other components also have the break points proportional to their atomic number, then the knee in the all-particle spectrum can be interpreted as the break point of the iron component. Extensive study of possible systematic errors are under way by both groups. It is still too early to draw any conclusions. For CORSIKA applications that are sensitive to low-energy muon numbers and energy spectra, the replacement of GHEISHA by FLUKA is recommended. New data for tuning the low-energy atmospheric neutrino simulation codes will be available soon from the NA49 and HARP groups. New projects, such as ARGO-YBJ, ASHRA-1, IceTop, KASCADE-Grande, 295

20 296 etc. will present interesting cosmic-ray data at ICRC2005 hopefully. 7. Acknowledgments The author would like to acknowledge valuable discussions with A. Haungs, G. Schatz, M. Roth, T. Yuda, M. Shibata, R. Engel, T. Montaruli and S. Ogio. He would also like to thank J. Huang, K. Kawata, S. Ozawa, A. Shiomi, H. Tsuchiya, S. Udo, and Y. Watanabe for help with figures. 1. Aglietta M. et al. 1999, Astropart. Phys., Aielli G. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Aita Y. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Amenomori M. et al. 1996, ApJ 461, Amenomori M. et al. 2000, Phys. Rev. D62, Amenomori M. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Amenomori M. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Antonov et al. 1999, Proc. of 26th ICRC (Utah), vol. 1, p Antonov R. A. et al. 1995, Proc of 24th ICRC (Rome), vol. 2, p Apanasenko A. V. et al. 2001, Astropart. Phys. 16, Arai Y. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Arqueros F. et al. 2000, Astron. Astrophys. 359, Asakimori K. et al. 1995, Proc. of 24th ICRC (Rome), vol. 2, p Barr G. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 3, p Celio P. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Chiavassa A. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Fomin Y. A. et al. 1991, Proc. of 22nd ICRC (Dublin), vol. 2, p Gaisser T. K. et al Proc. of 27th ICRC (Hamburg), vol. 5, p Gaisser T. K. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Glasmacher M. A. K. et al. 1999, Astropart. Phys., Gress O. A. et al. 1997, Proc. of 25th ICRC (Durban), vol. 4, p Grigorov N. L. et al. 1971, Proc. of 12th ICRC (Hobart), vol. 5, p Haungs A. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p He H. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Heck D. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Horneffer A. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Iacovacci M. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Ivanenko I. P. et al. 1993, Proc. of 23rd ICRC (Calgary), vol. 2, p Kopenkin V. V. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 3, p Kopenkin V. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 3, p

21 31. Kulikov G. V. et al. 1958, J. Exp. Theor. Phys. 35, Mastroianni S. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 2, p Mueller M. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Nagano M. et al. 1984, J. Phys. G: Nucl. Part. Phys., Ogio S. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Rawlins K. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Robbins S. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 3, p Roth M. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Shirk E. K. et al. 1978, ApJ. 220, Surdo A. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Sveshnikova L. G. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Swordy S. P. et al. 2000, Astropart. Phys. 13, Tokuno H. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p Valchierotti S. et al. 2003, Proc. of 28th ICRC (Tsukuba), vol. 1, p