1. LHCf: an accelerator experiment for Cosmic Ray Physics
3rd International Workshop on QCD at Cosmic Energies
Trieste May June 1, 2007
LHCf: an accelerator experiment for Cosmic Ray Physics
Oscar Adriani
University and INFN Firenze
on behalf of the LHCf Collaboration

2. The LHCf Collaboration
CERN
D.Macina, A.L. Perrot
USA
LBNL Berkeley:
W. Turner
FRANCE
Ecole Politechnique Paris:
M. Haguenauer
SPAIN
IFIC Valencia:
A.Fauss, J.Velasco
JAPAN:
STE Laboratory Nagoya University:
Y.Itow, T.Mase, K.Masuda,Y.Matsubara, H.Matsumoto,H.Menjo,T.Sako,H. Watanabe
Waseda University: K. Kasahara, Y.Shimizu, S.Torii
Konan University: Y.Muraki
Kanagawa University Yokohama: T.Tamura
ITALY
Firenze University and INFN: O.Adriani,, L.Bonechi, M.Bongi, G.Castellini, R.D’Alessandro, P.Papini
Catania University and INFN: A.Tricomi
Trieste, May 28, QCD at Cosmic Energies
LHCf

3. Outline Introduction Idea of LHCf Simulation and test beam results
Main problems in HECR physics: GZK cut off, chemical composition
Idea of LHCf
Measurement of neutral particles emitted very forward at LHC
Simulation and test beam results
Performances of the detector
Conclusions, plans and schedule
Toward the LHC operation
Trieste, May 28, QCD at Cosmic Energies
LHCf

4. Introduction: GZK cut off
GZK cutoff: 1020 eV
super GZK events?!?
Huge experiment (Auger, TA) will solve the statistics
29th ICRC
Pune (India)
AGASA reports 18% systematic uncertainty in energy determination.
10% of systematic is due to interaction model.
Accelerator calibration is necessary.
HOWEVER
Different interaction models give different answers for
the primary cosmic ray energy estimate.
20% correction on the
absolute energy scale!!!
Trieste, May 28, QCD at Cosmic Energies
LHCf

5. Introduction: cosmic ray composition
IRON
PROTON
Xmax(g/cm2)
LHCf
Not only GZK…
Different interaction models lead
different conclusions about the
composition of the primary cosmic rays.
UA7
Energy (eV)
Knapp et al., 2003
Trieste, May 28, QCD at Cosmic Energies
LHCf

6. Simulation plays a crucial role
Composition: inferred from Xmax
Spectrum: Energy is measured
by counting the secondaries
Simulation plays a crucial role
LHCf is a tool to calibrate the simulation
Trieste, May 28, QCD at Cosmic Energies
LHCf

7. Development of atmospheric showers
Simulation of an atmospheric shower due to a 1019 eV proton.
The dominant contribution to the energy flux is in the very forward region (  0)
In this forward region the highest energy available measurements of p0 cross section were done by UA7 (E=1014 eV, y= 5÷7)
Trieste, May 28, QCD at Cosmic Energies
LHCf

8. Longitudinal development of showers
Factor 2 of discrepancy
DPMJET, QGSJET, SIBYLL ...
are normally used
Sea Level
The direct measurement of the p production cross section as function of pT is essential to correctly estimate the energy of the primary cosmic rays
Trieste, May 28, QCD at Cosmic Energies
LHCf

9. Summarizing… LHC-HECR interplay LHC Calibration of the models at
high energy is mandatory
We will use LHC,
the highest energy accelerator
7 TeV + 7 TeV protons
14 TeV in the center of mass
Elab=1017 eV (Elab= E2cm/2 mP)
Major LHC detectors (ATLAS, CMS, LHCB) will measure
the particles emitted in the central region
LHCf will cover the very forward part
May be also heavy ions collisions????
LHC
Trieste, May 28, QCD at Cosmic Energies
LHCf

10. The LHC ring 27 km ring IP1 ATLAS
Trieste, May 28, QCD at Cosmic Energies
LHCf

11. LHCf: location and experimental layout
INTERACTION POINT
IP1 (ATLAS)
Beam line
Detector II
Tungsten
Scintillator
Silicon mstrips
Detector I
Scintillating fibers
140 m
Detectors should measure energy and position of g
from p0 decays e.m. calorimeters with
position sensitive layers
Two independent detectors on both side of IP1
Redundancy
Background rejection (especially beam-gas)
Trieste, May 28, QCD at Cosmic Energies
LHCf

12. LHCf location
Detectors will be installed in the TAN region, 140 m away from the Interaction Point, in front of luminosity monitors
Here the beam pipe splits in 2 separate tubes.
Charged particle are swept away by magnets!!!
We will cover up to y
Trieste, May 28, QCD at Cosmic Energies
LHCf

13. The TAN and LHCf manipulator marble shielding boxes for DAQ electronic
Trieste, May 28, QCD at Cosmic Energies
LHCf

14. Detector #1 Impact point (h) Energy
2 towers ~24 cm long stacked vertically with 5 mm gap
Lower:2 cm x 2 cm area
Upper: 4 cm x 4 cm area
4 pairs of scintillating fiber layers for tracking purpose (6, 10, 30, 42 r.l.)
Absorber
22 tungsten layers 7mm thick  44 X0 (1.6 lI) in total
(W: X0 = 3.5mm, RM = 9mm)
q < 300 mrad
16 scintillator layers (3 mm thick)
Trigger and energy profile measurements
Energy
Trieste, May 28, QCD at Cosmic Energies
LHCf

15. Transverse projection of detector #1 in the TAN slot
Trieste, May 28, QCD at Cosmic Energies
LHCf

16. Detector # 2
We use LHC style
electronics and readout
2 towers 24 cm long stacked on their edges and offset from one another
Lower:2.5 cm x 2.5 cm
Upper: 3.2 cm x 3.2 cm
4 pairs of silicon microstrip layers
(6, 12, 30, 42 r.l.) for tracking purpose (X and Y)  impact point
q < 400 mrad
16 scintillator layers (3 mm thick)
Trigger and energy profile measurements
Absorber
22 tungsten layers 7mm thick  44 X0 (1.6 lI) in total
(W: X0 = 3.5mm, RM = 9mm)
Energy
Trieste, May 28, QCD at Cosmic Energies
LHCf

17. Transverse projection of detector #2 in the TAN slot
Maximization of the acceptance in R (distance from beam center)
Trieste, May 28, QCD at Cosmic Energies
LHCf

18. Arm 1
Arm1 was fully assembled in Japan in July 2006 (scintillators + fibers + Tungsten) and fully tested at CERN in August 2006 beam test
Trieste, May 28, QCD at Cosmic Energies
LHCf

19. Scintillating fibers readout (Arm1)
4 cm
2 cm
Clear Fiber
SciFi 1mm Sq.
Joint
SciFi Belts
1 MIP > 5 p.e.
Hamamatsu
64 ch (8x8)
8 dynode
MAPMT
VA32HDR14 chip from IDEAS
1 ms shaping time
Huge dynamic range (30 pC)
32 channels
MAPMT+FEC
Trieste, May 28, QCD at Cosmic Energies
LHCf

20. Arm 2
Arm2 was partially assembled in Florence in July 2006 and brought to CERN for the Beam Test of August 2006
Arm2 was fully assembled in Florence in April 2007
Trieste, May 28, QCD at Cosmic Energies
LHCf

21. Silicon mstrips readout
Pace3 chips
(many thanks to CMS preshower!!!!)
32 channels
25 ns peaking time
High dynamic range (> 400 MIP)
192x32 analog pipeline
Trieste, May 28, QCD at Cosmic Energies
LHCf

22. The mechanics of the module
25 mm
Light Guides +
Scintillators
Hybrid circuit
Pace Chips
Silicon sensor
Kapton fanout
Tungsten
SiliconX SiliconY
Trieste, May 28, QCD at Cosmic Energies
LHCf

23. Trieste, May 28, 2007 - QCD at Cosmic Energies
LHCf

24. Front Counter (Both for Arm1&2)
read by a MAPMT
R7600U-00-M4
Trieste, May 28, QCD at Cosmic Energies
LHCf

25. Now the real installation in LHC…
The idea is that installation takes place in 2 steps:
Pre-installation
Final installation
In between the 2 installation the baking out of the beam pipe will be done (200 oC), so the detectors should be removed
Pre-Installation dates have been fixed in Fall 2006
LSS1L (Arm1)
8/01/2007 to 26/01/2007  FINISHED!
LSS1R (Arm2)
23/04/2007 to 11/05/2007  FINISHED
The dates for the final installation are still under discussion
Trieste, May 28, QCD at Cosmic Energies
LHCf

26. Arm1 pre-installation From 8/01/2007 to 26/01/2007
No major problems came out
Cables  OK
Transport and installation  OK
Laser calibration  OK
Power supply from USA15  OK
Manipulator and movements  OK
Arm1 was dismounted at the end
Trieste, May 28, QCD at Cosmic Energies
LHCf

27. Transport and insertion in the TAN
Trieste, May 28, QCD at Cosmic Energies
LHCf

28. LHCf Arm 1 – Installation completed within 15 minutes!
Trieste, May 28, QCD at Cosmic Energies
LHCf

29. Arm2 pre-installation From 23/04/2007 to 11/05/2007
No major problems came out
Cables  OK
Transport and installation  OK
Laser calibration  OK locally
 TBD from remote
Power supply from USA15  OK
Manipulator and movements  OK
No interference with BRAN
Arm2 was dismounted on May 9
Trieste, May 28, QCD at Cosmic Energies
LHCf

30. Trieste, May 28, 2007 - QCD at Cosmic Energies
LHCf

31. LHCf Physics performance
Single photon spectrum
p0 fully reconstructed (1 g in each tower)
p0 reconstruction is an important tool for energy calibration (p0 mass constraint)
Basic detector requirements:
minimum 2 towers (p0 reconstruction)
Smallest tower on the beam (multiple hits)
Dimension of the tower  Moliere radius
Maximum acceptance (given the LHC constraints)
Beam Test
Simulation
Trieste, May 28, QCD at Cosmic Energies
LHCf

32. Examples of simulated events for g and n
Trieste, May 28, QCD at Cosmic Energies
LHCf

33. LHCf performances: acceptance on PTg-Eg plane
Detectable events
140
Beam crossing
angle
A vertical beam crossing angle > 0 will
increase the acceptance of LHCf
Trieste, May 28, QCD at Cosmic Energies
LHCf

34. LHCf performances: single g geometrical acceptance
Some runs with LHCf vertically shifted few cm
will allow to cover the whole kinematical range
Trieste, May 28, QCD at Cosmic Energies
LHCf

35. LHCf performances: g shower in Arm #2
500 GeV g shower
Position resolution of detector # 2
Fluka based simulation
7 mm for 1.8 TeV g
Trieste, May 28, QCD at Cosmic Energies
LHCf

36. LHCf performances: p0 mass resolution
Arm #1
DE/E=5%
200 mm spatial resolution
Dm/m = 5%
Trieste, May 28, QCD at Cosmic Energies
LHCf

37. LHCf performances: Monte Carlo g-ray energy spectrum (5% Energy resolution is taken into account)
106 generated LHC interactions  1 Minute cm-2s-1 luminosity
Discrimination between various models
is feasible
Quantitative discrimination with the help of a properly defined c2 discriminating variable based on the spectrum shape
(see TDR for details)
Trieste, May 28, QCD at Cosmic Energies
LHCf

38. g ray energy spectrum for different positions
QGSJETII: used model
QGSJET: c2/DOF=107/125
DPMJET3: c2/DOF=224/125
SYBILL: c2/DOF=816/125
Trieste, May 28, QCD at Cosmic Energies
LHCf

39. LHCf performances: p0 geometrical acceptance
Arm #1
Arm #2
Trieste, May 28, QCD at Cosmic Energies
LHCf

40. LHCf performances: energy spectrum of p0
Typical energy resolution of g is 3 % at 1TeV
Trieste, May 28, QCD at Cosmic Energies
LHCf

41. LHCf performances: model dependence of neutron energy distribution
Original n energy
30% energy resolution
Trieste, May 28, QCD at Cosmic Energies
LHCf

42. Estimate of the background
beam-beam pipe
 E γ(signal) > 100 GeV, OK
background < 1%
beam-gas
 It depends on the beam condition
background < 1% (under Torr)
beam halo-beam pipe
 It has been newly estimated from the beam loss rate
Background < 10% (conservative value)
Trieste, May 28, QCD at Cosmic Energies
LHCf

43. The LPM effect Increase of X0 with increasing g energy
Landau – Pomeranchuk - Migdal
Increase of X0 with
increasing g energy
LHCf is able to directly
measure the LPM effect!
Trieste, May 28, QCD at Cosmic Energies
LHCf

44. SPS Beam Tests: 2004 & 2006 Setup Tests were successful
Energy resolution
Energy calibration
Spatial resolution of the tracking systems
CERN : SPS T2 H4
Incident Particles
Proton　 150,350GeV/c
Electron 100,200GeV/c
Muon GeV/c
Setup
LHCf
Detector
Silicon
Tracker
Moving Table
Trigger
Scintillator
Trieste, May 28, QCD at Cosmic Energies
LHCf

45. An electron event as seen by Arm1
40mm Calorimeter : 196 GeV electron
Transition Curve
Scifi e-
2 r.l. step
4 r.l. step
196GeV
Trieste, May 28, QCD at Cosmic Energies
LHCf

46. Leakage Correction Monte Carlo N Particles MC predicts that the
leakage is energy independent!
Prototype Experiment
Distance from Edge
correction
Trieste, May 28, QCD at Cosmic Energies
LHCf

47. Performances of the LHCf Detector
Measured at the SPS
Beam Test in 2004
SciFi Position Resolution
Energy Resolution
LHCf can measure (and provide to LHC) the center of neutral flux from the collisions
If the center of the
neutral flux hits LHCf
 << 1 mm resolution
Trieste, May 28, QCD at Cosmic Energies
LHCf

48. Energy Calibration
Problem: Determination of the conversion factor from ADC to Energy in each calorimeter layer
⇒　Ai [ADC counts/MeV] (i: Layer）
How to do?
Compare data with simulation (EPICS), tuning with beam test data
Event Selection (Centered events)
　　Incident point
　　　　20mm 3mm from calorimeter center
　　　　40mm 5mm from calorimeter center
10mm
6mm
Beam Profile
　40mm Calorimeter
　196 GeV e- RUN
Trieste, May 28, QCD at Cosmic Energies
LHCf

49. Data vs. Simulation Ai = Mean(data) / Mean(simulation)
40mm Calorimeter : 196 GeV 　Red : Simulation 　Black : Experiment
Ai = Mean(data) / Mean(simulation)
Layer χ2　 D.O.F.
/ 60
/ 48
/ 36
/ 38
Simulations and data agrees well!!!
Energy scale can be well inferred
Work in progress to check the energy scale for different energies
Trieste, May 28, QCD at Cosmic Energies
LHCf

50. Few plots of the beam test results for silicon
A high energy electron shower seen on x and y silicon X Y
<Noise> ~ 5 ADC counts
Trieste, May 28, QCD at Cosmic Energies
LHCf

51. Y Energy measured 100 GeV electrons X High Gain
Trieste, May 28, QCD at Cosmic Energies
LHCf

52. The LHCf operation in LHC

53. Optimal LHCf run conditions
Beam parameter
Value
# of bunches
≤ 43
Bunch separation
> 2 msec
Crossing angle
0 rad
140 mrad downward
Luminosity per bunch
< 2 x 1028 cm-2s-1
Luminosity
≤ ~1030 cm-2s-1
Bunch intensity
4x1010 ppb (b*=18 m)
1x1010 ppb (b*= 1 m)
Beam parameters used for commissioning are good for LHCf!!!
No radiation problem for 10kGy by a “year” operation with this luminosity
Trieste, May 28, QCD at Cosmic Energies
LHCf

54. From R. Bailey presentation at January 2007 TAN workshop
Optimal conditions for LHCf running!
Stage 1
Trieste, May 28, QCD at Cosmic Energies
LHCf

55. LHCf proposed running scenario
Phase-I
Run since the very beginning of LHC operations (Stage 1, 43 bunches)
Remove the detector for radiation issues when the machine goes to the Stage II (luminosity reaches 1031cm-2s-1) and reinstall the 3 Cu bars
Phase-II
Re-install the detector at the next opportunity of low luminosity run after removal of Cu bars (Totem dedicated runs? Possible LHCf dedicated runs?)
Phase-III
Future extension for p-A, A-A run with upgraded detectors?
Trieste, May 28, QCD at Cosmic Energies
LHCf

56. LHCf: conclusions and plans
LHCf approved in June 2006 by the LHCC
Physics performances:
able to measure π0 mass with 5% resolution.
able to distinguish the models by measurements of π0 and γ
able to distinguish the models by measurements of n
Detectors:
Arm1 & Arm2 are fully ready
Test beams were done in 2004 & 2006 to measure the performances
Installation phase well advanced
ARM1 already successfully pre-installed in January 07,
ARM2 already successfully pre-installed in April-May 07
Final installation date is under discussion
Running conditions:
Three foreseen phases
Phase I: Run at the beginning of LHC operations (43 bunches)
Phase II: operation during low luminosity TOTEM runs or dedicated runs
Phase III: Heavy Ion runs ?
Waiting for LHC beams!!!
Trieste, May 28, QCD at Cosmic Energies
LHCf

57. Back up Slides

58. Epics v.s. Geant4 e- W: 7mm x 16 Scintillator: 3mm 100mm
Trieste, May 28, QCD at Cosmic Energies
LHCf

59. Y X Beam profile 200 GeV electrons
Trieste, May 28, QCD at Cosmic Energies
LHCf

60. g rate
Trieste, May 28, QCD at Cosmic Energies
LHCf

61. p0 rate
Trieste, May 28, QCD at Cosmic Energies
LHCf

62. Effect of LHCf on BRAN measurement
LUMI monitor (BRAN) inside TAN is beyond LHCf (replacing 4th copper bar)
IP1
Cu Bar / ZDC
LHCf
LHCf
Cu Bar / ZDC
Lumi
Lumi
The effect of LHCf on BRAN measurements has been studied in the last months by simulation
Reduction of shower particles at BRAN
Position dependence on beam displacement
(question from machine peoples: if we shift by 1 mm the real beam, does the center of the measured neutral energy shifts by 1 mm?)
Trieste, May 28, QCD at Cosmic Energies
LHCf

63. BRAN response vs beam position
Relative change of the reduction factors for BRAN with respect to the nominal value (center of the beam: nominal one)
If the position of beam center stays within a few mm from the beam-pipe center, the reduction factors do not change more than 10%
1 x 1
cm2
Arm #1
Arm #2
H.Menjo
Trieste, May 28, QCD at Cosmic Energies
LHCf