Recent Results from the new Tibet hybrid experiment J. Huang for the Tibet ASγCollaboration a a Institute of high energy physics, Chinese Academy of Sciences China, Beijing 100049 a Today, I would like to talk about “Recent Results from the new Tibet hybrid experimen” International Conference ( HESZ2015 ) , Nagoya University, Japan

Contents Knee of the spectrum. Two Possible explanations for the sharp knee. New hybrid experiment (YAC+Tibet-III). Test of hadronic interaction models and compositions models. Primary proton, helium spectra obtained by (YAC-I + Tibet-III). Expected primary P, He and iron spectra by YAC-II Summary In this talk I would like to briefly summarize “Knee of the spectrum”, and “Two Possible explanations for the sharp knee”, and then I would like to introduce our “new hybrid experiment (YAC+ Tibet-III)”, and “Test of hadronic interaction models and compositions models”, and then “Primary proton and helium spectra obtained by (YAC-I + Tibet-III) experiment” will be discussed with some details. Also I would like to talk about “Expected results by (YAC-II + Tibet-III) “ and “Summary”. J. Huang (HESZ 2015, Japan)

All-particle spectrum measured by Tibet-III array from 1014 ~1017eV (ApJ 678, 1165-1179 (2008)) Model Knee Position (PeV) Index of spectrum QGS.+HD 4.0± 0.1 R1= -2.67± 0.01 R2= -3.10± 0.01 QGS.+PD 3.8± 0.1 R1= -2.65± 0.01 R2= -3.08± 0.01 SIB.+HD R2= -3.12± 0.01 The result of our measurement of all-particle energy spectrum covering three decades of energy range around the knee is shown in this Figure. The measurements in the highest energy range is also shown in this figure. 3 J. Huang (HESZ 2015, Japan) 5/ 31 3 3

Energy spectrum around the knee measured by many experiments 　　Tibet　　　　　　　　　　　　KASCADE　　　　　　　　HEGRA Energy spectrum around the knee measured by many experiments are also shown in these figures. CASA/MIA　　　　　　　　BASJE　　　　　　　　　　　Akeno DICE J. Huang (HESZ 2015, Japan)

Normalized spectrum J. Huang (HESZ 2015, Japan) When we normalized the all-particle energy spectrum at 10^15 eV, we found a sharp knee is clearly seen. J. Huang (HESZ 2015, Japan)

A sharp knee is clearly seen ( see the paper: M.Shibata, J. Huang et al. APJ 716 (2010) 1076 ) What is the origin of the sharp knee? There were many models: nearby source, new interaction threshold, etc. In the following, I would introduce our two analyses for the origin of the sharp knee. A sharp knee is clearly seen. Therefore we have a question, that is “what is the origin of the sharp knee ”, There were many models: nearby source, new interaction threshold, etc. In the following, we would introduce our two analyses for the origin of the sharp knee. 6 J. Huang (ISVHECRI2012, Berlin, Germany) 6/ 31 6

Two possible explanations for the sharp knee ( see the paper: M.Shibata, J. Huang et al. APJ 716 (2010) 1076 ) 2) Model B: Sharp knee is due to nonlinear effects in the diffusive shock acceleration (DSA) 1) Model A: Sharp knee is due to nearby sources In model A, sharp knee is due to nearby sources, in this model, and middle composition ( CNO ) is predicted at the knee. However, in Model B, sharp knee is due to nonlinear effects in the diffusive shock acceleration, in this model,and iron-dominant composition is predicted at the knee and beyond. All-particle knee = CNO? All-particle knee = Fe knee? J. Huang (HESZ 2015, Japan)

These aims will be realized by our new experiments 　　 In order to distinguish between Model A and Model B and many other models, measurements of the chemical composition around the knee, especially measurements of the spectra of individual component till their knee will be essentially important. Therefore, we planed a new experiment: 1) to lower down the energy measurement of individual component spectra to *10TeV and make connection with direct measurements; 2) to make a high precision measurement of primary p, He, …, Fe till 100 PeV region to see the rigidity cutoff effect. These aims will be realized by our new experiments YAC (Yangbajing AS Core array) ! J. Huang (HESZ 2015, Japan) 15/ 31

Knee of the spectrum The early works doing the research of the chemical composition of the knee region were the Tibet emulsion chamber (Tibet-EC) experiment and the KASCADE experiment. The results of the Tibet-EC suggest that the main component responsible for making the knee structure is composed of nuclei heavier than helium, however, the KASCADE claims that the knee is due to the steepening of the spectra of light elements with an exponential type cutoff as shown in this Figure. It is also noted that the experimental data points are still poor in the energy range between the direct measurements and the indirect measurements waiting us to study. Proton Helium The early works doing the research of the chemical composition of the knee region were the Tibet emulsion chamber (Tibet-EC) experiment and the KASCADE experiment. The results of the Tibet-EC suggest that the main component responsible for making the knee structure is composed of nuclei heavier than helium, however, the KASCADE claims that the knee is due to the steepening of the spectra of light elements with an exponential type cutoff as shown in this Figure. It is also noted that the experimental data points are still poor in the energy range between the direct measurements and the indirect measurements waiting us to study. The YAC-II aims to observe the energy spectrum of proton , helium, iron whose energy range will overlap with direct observations at lower energies such as CREAM, ATIC and TRACER, and Tibet-EC experiment at higher energies. J. Huang (HESZ 2015, Japan)

Knee of the spectrum The YAC-II aims to observe the energy Proton Helium YAC The YAC-II aims to observe the energy spectrum of proton , helium, iron whose energy range will overlap with direct observations at lower energies such as CREAM, ATIC and TRACER, and Tibet-EC experiment at higher energies. The early works doing the research of the chemical composition of the knee region were the Tibet emulsion chamber (Tibet-EC) experiment and the KASCADE experiment. The results of the Tibet-EC suggest that the main component responsible for making the knee structure is composed of nuclei heavier than helium, however, the KASCADE claims that the knee is due to the steepening of the spectra of light elements with an exponential type cutoff as shown in this Figure. It is also noted that the experimental data points are still poor in the energy range between the direct measurements and the indirect measurements waiting us to study. The YAC-II aims to observe the energy spectrum of proton , helium, iron whose energy range will overlap with direct observations at lower energies such as CREAM, ATIC and TRACER, and Tibet-EC experiment at higher energies. J. Huang (HESZ 2015, Japan)

New hybrid experiment (YAC-II+Tibet-III+MD) This hybrid experiment consists of low threshold Air shower core array (YAC-II) and Air Shower array ( Tibet-III ) and Muon Detector ( MD ) . Pb 7 r.l. Scint. Iron YAC-II aims to measure the primary energy spectrum of 4 mass groups of Proton, Helium, iron and 4<A<56, at 50 TeV – 1016 eV range covering the knee. MD AS YAC2 This new hybrid experiment consists of low threshold Burst Dector grid (YAC) and Air Shower (AS ) array and Muon Detector ( MD ) . Tibet AS array is used to measure primary energy and incident direction, and YAC2 is used to measure high energy AS core within several x 10m from the axis, and Tibet MD array is used to measure number of muon. This hybrid experiment (YAC+Tibet-AS + MD) aims to measure the primary energy spectrum of 4 mass groups of P, He, Iron and 4< A < 56 at 50 TeV – 1016 eV range covering the knee. Tibet-III (50000 m2) : Primary energy and incident direction. YAC-II ( 500 m2 ): High energy AS core within several x 10m from the axis. Tibet-MD ( 4500 m2 ) : Number of muon. J. Huang (HESZ 2015, Japan) 11

- Full M.C. Simulation - Hadronic interaction model CORSIKA (Ver. 7.3500 ) – EPOS -LHC– – QGSJETII-04– – SIBYLL 2.1 – 　　 = Air Shower simulation = CORSIKA 7.3500 (EPOS –LHC, QGSJETII-04, SIBYLL2.1) ( 1 ) Primary energy: E0 >1 TeV ( 2 ) All secondary particles are traced until their energies become 300 MeV in the atmosphere. ( 3 ) Observation Site : Yangbajing (606 g/cm2 ) 　　 = Detector simulation = Geant 4 ( Ver. 9.5) Simulated air-shower events are reconstructed with the same detector configuration and structure as the Tibet-III YAC and Muon detector array. Primary composition model 　Helium poor model [1] 　Helium rich model [1] Gaisser’s fit model [1] see [1] J. Huang et al. Astropart. Phy 66 (2015) 18 We have carried out a full Monte Carlo (MC) simulation on the development of air showers in the atmosphere using the simulation code Corsika (version 7.3500). Three hadronic interaction models, including SIBYLL2.1, EPOS-LHC (v3400)and QGSJETII-04, are used to generate the air-shower events in the atmosphere. For the primary cosmic rays, we examined three composition models, namely, ``He-poor", ``He-rich" and ``Gaisser-fit" models, in order to evaluate the systematic errors attributable to primary composition models. For the details, please see the paper [1] J. Huang et al. Astropart. Phy 66 (2015) 18 J. Huang (HESZ 2015, Japan) 20/ 28 12

Primary cosmic-ray composition models ( see the paper: J. Huang et al. Astropart. Phy. 66 (2015) 18 ) The main differences between “He-poor model” , “He-rich model” and “Gaisser’s fit model” are as follows: The He spectrum of “He-poor model” coincides with the results from RUNJOB, The He spectrum of “He-rich model” coincides with the results from CREAM, ATIC2 and JACEE. The “Gaisser-fit model” fits to a higher He model (almost same as our “He-rich model”) at the low energy range and to the KASCADE-QGSJET data at high energy range in which light components dominate in the chemical composition. In all models mentioned above, each component is summed up so as to match with the all-particle spectrum with a sharp knee, which was obtained with the Tibet-III AS array. The main differences between “He-poor model” , “He-rich model” and “Gaisser’s fit model” are as follows: The He spectrum of “He-poor model” coincides with the results from RUNJOB, The He spectrum of “He-rich model” coincides with the results from CREAM, ATIC2 and JACEE. The “Gaisser-fit model” fits to a higher He model (almost same as our “He-rich model”) at the low energy range and to the KASCADE-QGSJET data at high energy range in which light components dominate in the chemical composition. In all models mentioned above, each component is summed up so as to match with the all-particle spectrum with a sharp knee, which was obtained with the Tibet-III AS array. J. Huang (HESZ 2015, Japan)

Tests of hadronic interaction models by ( YAC-I + Tibet-III ) J. Huang (HESZ 2015, Japan)

Prototype experiment (YAC-I+Tibet-III) ( data taking started from 2009 Prototype experiment (YAC-I+Tibet-III) as shown in this Figure. Data taking started from 2009.04.01. YAC1 consists of 16 YAC detectors, covering an area about 10 m^2. Total : 16 YAC detectors Effective area: 10 m2 J. Huang (HESZ 2015, Japan) 19/ 31

Observable parameters of high energy cores as sumNb , Tests of hadronic interaction models and composition models (Some preliminary results from YAC-I data samples, primary energy: 100TeV -1PeV Observable parameters of high energy cores as sumNb , Nb_top, <R> and <Nb*R> are calculated for various combination of interaction and primary models and are compared each others as shown in Figure. The area of all distributions are normalized as their total area =1. The shape of each parameter distribution is almost same since these cores are mainly multiple EM cascade processes originated by decayed gamma rays from neutral pions. The experimental data are consistent with those, which mean that there exist no detection bias. Note that the shape shows a (cascade) scaling behavior, while its intensity depends on the primary model. Observable parameters of high energy cores as sumNb , Nb_top, <R> and <Nb*R> are calculated for various combination of interaction and primary models and are compared each others as shown in Figure. The area of all distributions are normalized as their total area =1. The shape of each parameter distribution is almost same since these cores are mainly multiple EM cascade processes originated by decayed gamma rays from neutral pions. The experimental data are consistent with those, which mean that there exist no detection bias. Note that the shape shows a (cascade) scaling behavior, while its intensity depends on the primary model. J. Huang (HESZ 2015, Japan)

preliminary preliminary Some discrepancies in the absolute intensities are seen: First, note that the three interaction models give same intensity when the primary is the “He-poor” model. This result suggest that the difference among three interaction models are very small, say at most 10% since almost all high energy cores are produced by same primary protons (He contribution is very small). Experimental Data are very close to SIBYLL2.1, EPOS-LHC and QGSJETII-04, if we assumed “He-rich” primary composition. It means that YAC-experimental data have the same primary composition with the “ He-rich” model. However, the “Gaisser’s fit” primary model always gives a higher flux by a factor 1.4 - 2.0 than the experimental result and other primary models. It means that there are too many primary light components (P,He) in “Gaisser’s fit” model than experimental data. These results well show that YAC is very sensitive to observe light components and well reflet the structure of primary spectrum of light components as seen in this figure. Comparison of event absolute intensities between Expt. and MC preliminary preliminary Expt.data For details, please see the following paper: L.M.Zhai, J.Huang et al., J.Phys. G: Nucl. Part. Phys. 42 (2015), 045201 Some discrepancies in the absolute intensities are seen: First, note that the three interaction models give same intensity when the primary is the “He-poor” model. This result suggest that the difference among three interaction models are very small, say at most 10% since almost all high energy cores are produced by same primary protons (He contribution is very small). Experimental Data are very close to SIBYLL2.1, EPOS-LHC and QGSJETII-04, if we assumed “He-rich” primary composition. It means that YAC-experimental data have the same primary composition with the “ He-rich” model. However, the “Gaisser’s fit” primary model always gives a higher flux by a factor 1.4 - 2.0 than the experimental result and other primary models. It means that there are too many primary light components (P,He) in “Gaisser’s fit” model than experimental data. These results well show that YAC is very sensitive to observe light components and well reflet the structure of primary spectrum of light components as seen in this figure. Fig. 1: Intergral burst-size spectra (SumNb) Fig. 2 : The intensity ratios of (SumNb) to that obtained by the Expt. Data. The red dash line (Ratio =1) is Expt.data. J. Huang (HESZ 2015, Japan)

Comparison of event mean lateral spreads (<NbR>) between Expt Comparison of event mean lateral spreads (<NbR>) between Expt. and MC preliminary Some discrepancies in the mean lateral spreads are seen: This conclusion is same as the above Fig.1 and Fig.2’s conclusion, that is, YAC-experimental data are very close to “EPOS-LHC + He-rich” and “QGSJETII-04 + He-poor ” model. However, the “Gaisser’s fit” primary composition model always gives a lower distribution. It means that there are too many primary light components (P,He) in “Gaisser’s fit” than experimental data. Mean lateral spreads of AS-cores events are compared with the MC simulation results. It is known that although these are not strong, it is dependent on the transverse momentum Of secondary particles produced in hadronic interactions as well as interaction height as seen in Fig.3. Some discrepancies in the mean lateral spreads are seen: This conclusion is same as the above Fig.1 and Fig.2’s conclusion, that is, YAC-experimental data are very close to “EPOS-LHC + He-rich” and “QGSJETII-04 + He-poor ” model. However, the “Gaisser’s fit” primary composition model always gives a lower distribution. It means that there are too many primary light components (P,He) in “Gaisser’s fit” than experimental data. Fig. 3. Mean lateral spread of AS-cores J. Huang (HESZ 2015, Japan)

Primary proton, helium spectra analysis Hereafter , I would like to report “Primary proton, helium spectra analysis”. J. Huang (HESZ 2015, Japan)

Primary proton, He spectra analysis Identification of proton events ANN (a feed-forward artificial neural network) is used. Input event features: Ne, ΣNb, Nbtop, Nhit, <Rb>, < NbRb>, θ Classification: proton/others Primary (proton+He) energy determination E0=f(Ne,s) based on (p+He)-like MC events In this work, the separation of the primary mass is realized by use of a feed-forward artificial neural network (ANN) method. The following 7 parameters are input to the ANN . Primary energy is determinate by using the correlation between shower size and primary cosmic ray energy Based on proton-like MC events. J. Huang (HESZ 2015, Japan)

Air shower size to primary energy The primary energy (E0 ) of each AS event is determined by the air-shower size (Ne) which is calculated by fitting the lateral particle density distribution to the modified NKG function. Modified NKG function Hereafter , I would like to introduce how can I get primary energy. We used J. Huang (HESZ 2015, Japan)

Air shower size to primary energy E0=1.88*( Ne/1000.0)0.92 ) 　(TeV) Energy resolution ~ 20% ( at 200 TeV ) This figure shows the correlation between shower size and primary cosmic ray energy. Based on this figure, we convert shower size to primary energy. One event energy resolution is estimated to be 20% approximately. J. Huang (HESZ 2015, Japan)

Primary (P+He) separation by ANN for MC events QGSJET Purity – 95.2% Efficiency – 62% P+He Other Nuclei SIBYLL Purity – 94.5% Efficiency – 63% Primary (P+He) separation by ANN for MC events as shown in this figure, together with the experimental results. One can see that, the experimental data is also in a good agreement with the MC prediction, and that the (P+He) induced events are clearly separated from other nuclei. When Tc value is set as 0.2, where the purity is about 95.2%, and the efficiency is about 62%. P+He Other Nuclei J. Huang (HESZ 2015, Japan)

Results and discussions   preliminary   The spectral index of primary (P+He) spectra is –2.66 below 100 TeV, while it is -3.04 above 400 TeV in the case of QGSJET+HD model, and –2.65 below 100 TeV, -3.07 above 400 TeV in the case of SIBYLL + HD model Various Systematic errors are under study now ! And the Knee position of primary (P+He) spectra is 410 ±49 TeV in the case of QGSJET+HD model, 397 ±43 TeV in the case of SIBYLL + HD model, J. Huang (HESZ 2015, Japan)

(YAC-II +Tibet-III+MD) is well running now ( data taking started from 2014. 2) Pb 50cm 80cm (YAC-II +Tibet-III+MD) is well running now, data taking started from 2014. 2 There are 124 YAC-II detectors, cover area is about 500 m^2. Total : 124 YAC detectors Cover area: ~ 500 m2 J. Huang (HESZ 2015, Japan) 16/ 28

Expected results by (YAC-II+Tibet-III +MD) MC simulation shows that (YAC-II+Tibet-III +MD) is powerful enough to obtain proton, helium, iron and 4<A<56 spectra of primary CRs in the range of 50 TeV -1016 eV with an energy resolution better than 12% at 1015 eV. Expected primary P, He spectra Expected primary Iron spectra Iron Expected results by (YAC-II+Tibet-III+MD) is shown in this figure for proton, helium and iron group. MC shows that this hybrid experiment has a capability of measuring the primary energy spectrum P, He, Iron at 50 TeV -10 PeV range covering the knee. 26 J. Huang (HESZ 2015, Japan) 26

Summary 1. YAC-I shows the ability and sensitivity in checking the hadronic interaction models. Based on the “He-poor” primary model, we estimate that the difference of EPOS-LHC, QGSJETII-04, SIBYLL2.1 is within 10 % in our concerned energy region. High core events are very sensitive to the light components in CRs and the core parameters of sum Nb, Nb_top, <R> and <Nb*R> are very useful to separate the light components from all the observed events using a ANN technique. The flux of high energy core events are sensitive to the light components, since the interaction model seems to depend strongly on the production of high energy core events, we can obtain the energy spectrum of light components in primary CRs with sufficient accuracy as discussed in the YAC-I experiment. Simultaneous observation of core events with YAC and MD array may allow us to obtain the spectrum of each of primary components separately. Actually, the full MC simulation shows that the (YAC-II+Tibet-III +MD) array is powerful enough to obtain the energy spectra of proton, helium, medium nuclei and iron spectra of primary CRs in the range of 50 TeV -10 PeV with sufficient accuracy. It is estimated that the energy resolution of primary particles of 1 PeV is better than 12%。

Summary 4. We obtained the primary energy spectrum of proton, helium and (P+He) spectra between 50 TeV and 1 PeV , and found that the knee of the (P+He) spectra is located around 400 TeV. 5. The interaction model dependence in deriving the primary proton, helium and (P+He) spectra are found to be small (less than 25% in absolute intensity, 10% in position of the knee ), and the composition model dependence is less than 10% in absolute intensity, and various systematic errors are under study now ! 6. Next phase experiment YAC2 will measure the primary energy spectrum of 4 mass groups of P, He, 4<A<40, A>40 at 1014 – 1016 eV range covering the knee.

Thank you for your attention !!

Core event selection Selected Events QGSJET+HD 216942 SIBYLL+HD 304785 Event selection condition for AS core event was studied by MC and following criteria were adopted to reject non core events whose shower axis is far from the YAC array. Nb>200, Nhit≧4, Nbtop ≧1500, Ne>80000 | AS axis by LDF – burst center| < 5 m Statistics of core events in MC simulation and experiment Live Time is 106.05 days. Selected Events QGSJET+HD 216942 SIBYLL+HD 304785 QGSJET+NLA 80861 SIBYLL+NLA 64331 YAC1 5035 Event selection condition for AS core event was studied by MC and following criteria were adopted to reject non core events whose shower axis is far from the YAC array. Nb>200, Nhit≧4, Nbtop ≧1600, Ne>80000 | AS axis by LDF – burst center| < 5 m Statistics of core events in MC simulation and experiment are shown in this Table. J. Huang (ISVHECRI2012, Berlin, Germany)

measure the P,He and iron component at the knee Sensitivity of (YAC-II+Tibet-III+MD) to measure the P,He and iron component at the knee The above mentioned parameters observed with the (Tibet-III +YAC-II + MD ) array are very sensitive to different primary CR components. Used in this work is a feed-forward artificial neural network ( ANN) method to separate P-like events, He-like events and Iron-like events by using the above parameters. MC shows that this hybrid experiment has a capability of measuring the primary energy spectrum P, He, Iron at 50 TeV -10 PeV range covering the knee. Other like Iron like Purity – 83.5% Efficiency – 65.3%