1. COSMIC-RAY PHYSICS AT LHC: LATEST RESULTS FROM THE LHCf EXPERIMENT 24 th European Cosmic Ray Symposium 2014 Kiel, 3 rd Sep 2014 Massimo Bongi University of Florence and INFN

2. M. Bongi – ECRS2014 – 3 rd Sep 2014 OUTLINE 2

3. Extensive air shower observation longitudinal distribution lateral distribution arrival direction Astrophysical parameters spectrum composition source distribution (shower development) X max : depth of shower maximum in the atmosphere gives information on the CR composition Uncertainty of hadron interaction models Uncertainty in the interpretation of HIGH ENERGY COSMIC RAYS 3 PROTON IRON

4. M. Bongi – ECRS2014 – 3 rd Sep 2014 ① Inelastic cross section If large  rapid development If small  deep penetrating ② Forward energy spectrum If softer: rapid development If harder: deep penetrating If large k (  0 s carry more energy): rapid development If small k (baryons carry more energy): deep penetrating HOW ACCELERATOR EXPERIMENTS CAN CONTRIBUTE 4 ④ Nuclear effects CR-Air

5. M. Bongi – ECRS2014 – 3 rd Sep 2014 TUNING OF HADRON INTERACTION MODELS AFTER THE FIRST LHC DATA Significant reduction of differences among hadronic interaction models Courtesy of T. Pierog Karlsruhe Institute of Technology, Karlsruhe, Germany before LHCafter LHC 5

6. M. Bongi – ECRS2014 – 3 rd Sep 2014 THE LHCf COLLABORATION 6

7. M. Bongi – ECRS2014 – 3 rd Sep 2014 THE LHCf COLLABORATION 7

8. M. Bongi – ECRS2014 – 3 rd Sep 2014 LHCf EXPERIMENTAL SET-UP 8 140 m n π0π0 γ γ 8 cm Arm1 6 cm Arm2 IP 1 single beam pipe two beam pipes

9. M. Bongi – ECRS2014 – 3 rd Sep 2014 40mm 20mm ‐Energy resolution (> 100GeV) < 5% for  and  (35-40)% for n ‐Position resolution < 200μm (Arm1) and  40μm (Arm2) Sampling electromagnetic calorimeters  Absorber: 22 tungsten layers (total depth: 44 X 0, 1.6 λ I )  Energy measurement: 3-mm thick plastic scintillator tiles  4 tracking double (XY) layers for position reconstruction: SciFi (Arm1) and Silicon  -strip (Arm2)  Each detector has two independent calorimeter towers  reconstruction of  0   events thin scintillators 80x80 mm 2 upstream the calorimeters monitoring of beam condition background rejection luminosity estimation Arm1 Front Counters Arm2 25mm 32mm LHCf DETECTORS AND PERFORMANCES 9 Pseudo-rapidity  =-ln(tan(θ/2)):  > 8.7 @ zero X-ing angle  > 8.4 @ 140-  rad X-ing angle Performance beam

10. M. Bongi – ECRS2014 – 3 rd Sep 2014 SOME PICTURES OF LHCf 10

11. M. Bongi – ECRS2014 – 3 rd Sep 2014 PHYSICS PROGRAM AND TABLE OF PUBLICATIONS 11 Int. J. Mod. Phys. A 28, 1330036 (2013)

12. M. Bongi – ECRS2014 – 3 rd Sep 2014 12

13. M. Bongi – ECRS2014 – 3 rd Sep 2014 THE CHALLENGE OF NEUTRON ANALYSIS  Performance for neutrons is estimated from MC & test beams to be worse than for photons:  E /E ~ (35  40) %  x ~ (0.2  1) mm  In addition, detector performance shows some dependence on the hadron interaction model used for the calibration of the detector response  Unfolding is essential to extract physics results from the measured spectra  Measurement of neutron component is important to try to solve the muon excess observed by ground based HECR experiments 13

14. M. Bongi – ECRS2014 – 3 rd Sep 2014 For  <9.22 predictions by DPMJET and PYTHIA models show partial agreement with data Very large peak at high energy in the range  >10.76 (predicted only by QGSJET)  small inelasticity in the very-forward region INCLUSIVE NEUTRON SPECTRA (p-p@7TeV) 14 Before unfolding After unfolding Preliminary

15. M. Bongi – ECRS2014 – 3 rd Sep 2014 proton impact parameter b protonPb Central collisions (soft) QCD: central and peripheral collisions Ultra Peripheral Collisions (UPC): collisions with virtual photons from rel. Pb Dominant channel to forward  0 is:  About half of the observed  0 s originate from UPC  About half is from soft-QCD  Need to subtract UPC component Comparison with soft-QCD  0 ANALYSIS IN p-Pb COLLISIONS @ 5.02 TeV 15 Peripheral collisions Estimation of momentum distribution of the UPC induced secondary particles (Lab frame+Boost): 1.energy distribution of virtual photons is estimated by the Weizsacker-Williams approximation 2.photon-proton collisions are simulated by the SOPHIA model (E γ > pion threshold)

16. M. Bongi – ECRS2014 – 3 rd Sep 2014  0 EVENT RECONSTRUCTION IN p-Pb COLLISIONS 16 1) Search for two photons2) BG subtraction by sideband 4) Subtraction of the UPC component 3) Unfolding of the smeared p T spectra and correction for geometrical inefficiency π 0 detection efficiency p-Pb √s=5.02TeV (PROTON-REMNANT SIDE)

17. M. Bongi – ECRS2014 – 3 rd Sep 2014 INCLUSIVE  0 p T -SPECTRA IN p-Pb @ 5.02 TeV 17 LHCf p-Pb data (filled circles) show good agreement with DPMJET and EPOS. LHCf p-Pb spectra are harder than LHCf “p-p@5.02TeV” data (shaded area, multiplied by 5). The latter is interpolated from the results of p-p@2.76TeV and p-p@7TeV.

18. M. Bongi – ECRS2014 – 3 rd Sep 2014 NUCLEAR MODIFICATION FACTOR IN p-Pb @ 5.02 TeV 18  Both LHCf and MC data show strong suppression.  NMF grows with increasing p T, as can be expected by the p T spectrum that is softer in “p-p @ 5.02 TeV” than in p-Pb @ 5.02 TeV collisions.

19. M. Bongi – ECRS2014 – 3 rd Sep 2014 DETECTOR UPGRADE TO COMPLETE THE MAIN PHYSICS PROGRAM 19

20. M. Bongi – ECRS2014 – 3 rd Sep 2014 NEW POSSIBLE RUN AT RHIC: THE RHIC f PROGRAM 20 Preliminary

21. M. Bongi – ECRS2014 – 3 rd Sep 2014 CONCLUSIONS 21

22. M. Bongi – ECRS2014 – 3 rd Sep 2014 BACKUP SLIDES 22

23. M. Bongi – ECRS2014 – 3 rd Sep 2014 CALIBRATION OF HADRON INTERACTION MODELS AT LHC 23 p-p 450 GeV + 450 GeV  E lab ~ 4  10 14 eV p-p 3.5 TeV + 3.5 TeV  E lab ~ 3  10 16 eV p-p 6.5 TeV + 6.5 TeV  E lab ~ 9  10 16 eV  Total cross section ↔ TOTEM, ATLAS, CMS  Multiplicity ↔ Central detectors  Inelasticity/Secondary spectra ↔ Forward calorimeters ( LHCf, ZDCs) Multiplicity @ 14TeVEnergy Flux @ 14TeV (simulated by DPMJET3)

24. M. Bongi – ECRS2014 – 3 rd Sep 2014 Determination of energy from total energy release PID from shape Determination of the impact point Measurement of the opening angle of gamma pairs Identification of multiple hit 25mm Tower32mm Tower  600GeV photon  420GeV photon Longitudinal development measured by scintillator layers Transverse profile measured by silicon  –strip layers ` X-view Y-view ` Reconstruction of  0 mass: A VERY CLEAR  0 IN ARM2 24

25. M. Bongi – ECRS2014 – 3 rd Sep 2014 PARTICLE IDENTIFICATION 25 500 GeV < E REC < 1 TeV photon hadron 44 X 0 1.55 λ  L 90% : longitudinal position containing 90% of the shower energy  Photon selection based on L 90% cut  Energy dependent threshold in order to keep constant efficiency ε PID = 90%  Purity P = N phot /(N phot +N had ) estimated by comparison with MC  Event number in each bin corrected by P/ε PID  MC photon and hadron events are independently normalized to data  Comparison done in each energy bin  LPM effect is switched on

26. M. Bongi – ECRS2014 – 3 rd Sep 2014 Syst.+Stat. DATA DPMJET 3.04 QGSJET II-03 SIBYLL 2.1 EPOS 1.99 PYTHIA 8.145 No strong evidence of  -dependence SYBILL and EPOS show reasonable agreement of shape None of the models reproduces LHCf data within the error bars 26

27. M. Bongi – ECRS2014 – 3 rd Sep 2014 No model can reproduce the LHCf data perfectly DPMJET and PYTHIA are in good agreement at high-η for 0.5 1.5TeV QGSJET, SIBYLL, EPOS show reasonable agreement of shape for high η, but not for low η Syst.+Stat. DATA DPMJET 3.04 QGSJET II-03 SIBYLL 2.1 EPOS 1.99 PYTHIA 8.145 27

28. M. Bongi – ECRS2014 – 3 rd Sep 2014 DATA VS MC: COMPARISON 900 GeV / 7 TeV 28 900GeV 7TeV η>10.948.81<η<8.9 None of the model shows a complete agreement with the LHCf data up to x2 factor differences

29. M. Bongi – ECRS2014 – 3 rd Sep 2014 EPOS shows the best agreement with data DPMJET and PYTHIA have harder spectra than data QGSJET has softer spectrum than data 29

30. M. Bongi – ECRS2014 – 3 rd Sep 2014  0 p T -SPECTRA FOR DIFFERENT y RANGES: MC/DATA 30 EPOS gives the best agreement both for shape and yield. DPMJET 3.04 QGSJETII-03 SIBYLL 2.1 EPOS 1.99 PYTHIA 8.145 00.6P T [GeV] 00.6P T [GeV]00.6P T [GeV]00.6P T [GeV] 00.6P T [GeV]00.6P T [GeV] MC/Data

31. M. Bongi – ECRS2014 – 3 rd Sep 2014 L 90% L 20% Layer[r.l.] hadron photon projection along the sloped line L 90% L 20% Shower development in the small calorimeter tower NEUTRON IDENTIFICATION Particle Identification with high efficiency and small contamination is necessary A 2D method based on longitudinal shower development is used L 20% (L 90% ): depth in X 0 where 20% (90%) of the deposited energy is contained L 2D = L 90% - 0.25 * L 20% Mean purity in the 0-10 TeV range: 95% Mean efficiency: ~90% L 2D 31

32. M. Bongi – ECRS2014 – 3 rd Sep 2014 MUON EXCESS AT PIERRE AUGER OBS. 32 Pierre Auger Collaboration, ICRC 2011 (arXiv:1107.4804) Pierog and Werner, PRL 101 (2008) 171101 Auger hybrid analysis (QGSJETII.03) event-by-event MC selection to reproduce FD data (longitudinal profile) comparison with SD data (lateral profile) muon excess in data even for Iron primary MC EPOS predicts more muons due to larger baryon production => importance of baryon measurement

33. M. Bongi – ECRS2014 – 3 rd Sep 2014 Motivations: Inelasticity measurement: k=1-p leading /p beam Muon excess at Pierre Auger Observatory c osmic rays experiment measure HECR energy from muon number at ground and florescence light 20-100% more muons than expected have been observed Number of muons depends on the energy fraction of produced hadron Muon excess in data even for Fe primary MC EPOS predicts more muon due to larger baryon production R. Engel importance of baryon measurement THE IMPORTANCE OF NEUTRONS IN THE VERY FORWARD REGION 33

34. M. Bongi – ECRS2014 – 3 rd Sep 2014  2013 Jan-Feb: p-Pb/Pb-p collisions Installation of only Arm2 at one side (Silicon tracker good for high multiplicity) Data taken both at p-side (20 Jan – 1 Feb) and Pb- side (4 Feb), thanks to the swap of the beams  Details of beams and DAQ L = 1x10 29 – 0.5x10 29 cm -2 s -1 ~200.10 6 events  * = 0.8 m, 290  rad crossing angle 338p+338Pb bunches (min.  T = 200 ns), 296 colliding at IP1 10-20 kHz trigger rate downscaled to approximately 700 Hz 20-40 Hz ATLAS common trigger. Coincidence successful! p-p collisions at 2.76 TeV have also been taken p Pb IP8 IP2 IP1 Arm2 34

35. M. Bongi – ECRS2014 – 3 rd Sep 2014 PRELIMINARY p-Pb RUN:  0 35

36. M. Bongi – ECRS2014 – 3 rd Sep 2014 UPC SUBTRACTION 36

37. M. Bongi – ECRS2014 – 3 rd Sep 2014 DERIVATION OF  0 p T -SPECTRA IN p-p @ 5.02 TeV 37 1. Thermodynamics (Hagedorn model) 2. Gauss distribution The p T spectra for “ p-p @ 5.02 TeV ” are obtained by the Gauss distribution with the interpolated and absolute normalization.

38. M. Bongi – ECRS2014 – 3 rd Sep 2014 COMPARISON OF DATA: 900 GeV vs 7 TeV 38 Preliminary Data at √s=900GeV (normalized by the number of entries in X F > 0.1) Data 2010 at √s=7TeV (η>10.94) Normalized by the number of entries in X F > 0.1 No systematic error is considered in both collision energies. X F spectra: 900GeV data vs 7TeV data Coverage of 900GeV and 7TeV results in P T and Feynman-X Xf=E/Ebeam Good agreement of X F spectrum shape between 900 GeV and 7 TeV.  weak dependence of on E CMS

39. M. Bongi – ECRS2014 – 3 rd Sep 2014 39

40. M. Bongi – ECRS2014 – 3 rd Sep 2014 IDEA FOR THE REDUCTION OF ELECTRONICS SATURATION 40

41. M. Bongi – ECRS2014 – 3 rd Sep 2014 PHYSICS OF RHICf 41  Physics of RHICf  Energy Scaling of very-forward region in p-p collisions at √s=500GeV  Measurement at p – light-ion collisions (p-O) √s NN =200GeV  Asymmetry of Forward Neutron with polarized beams  LOI submitted to the RHIC committee and nicely appreciated The STAR Collaboration, PRL 97 (2006) 152302 Nuclear modification factor in d-Au collisions at √s NN =200 GeV