13th ICATPP Conference on
Astroparticle, Particle, Space Physics and Detectors for Physics Applications
Villa Olmo - Como, Oct 3rd – 7th, 2011
Recent results on neutral
particles spectra from the
LHCf experiment
Massimo Bongi - INFN (Florence, Italy)
LHCf Collaboration
High-energy cosmic rays
SPS
Tevatron
LHC
Recent excellent observations (e.g.
Auger, HiRes, TA) but the origin and
composition of HE CR is still unclear
AUGER
E [eV]
Massimo Bongi – ICATPP – 3rd October 2011 – Como
2
Development of atmospheric showers
 The depth of the maximum of the shower Xmax in the atmosphere depends on
energy and type of the primary particle
 Several Monte Carlo simulations (different hadronic interaction models) are used
and they give different answers about composition
1019 eV proton
Experimental tests of hadron interaction models are
necessary
The dominant contribution to the shower
development comes from particles emitted at
low angles (forward region).
LHC gives us the unique opportunity to study hadronic
interactions at 1017eV
7 TeV + 7 TeV
→ Elab ≈ 1 x 1017 eV
3.5 TeV + 3.5 TeV → Elab ≈ 3 x 1016 eV
450 GeV + 450 GeV → Elab ≈ 4 x 1014 eV
LHC forward (LHCf) experiment
Massimo Bongi – ICATPP – 3rd October 2011 – Como
3
The LHCf collaboration
K.Fukatsu, T.Iso, Y.Itow, K.Kawade, T.Mase, K.Masuda,
Y.Matsubara, G.Mitsuka, Y.Muraki, T.Sako, K.Suzuki, K.Taki
Solar-Terrestrial Environment Laboratory, Nagoya University, Japan
H.Menjo
Kobayashi-Maskawa Institute, Nagoya University, Japan
K.Yoshida
Shibaura Institute of Technology, Japan
K.Kasahara, T.Suzuki, S.Torii
Waseda University, Japan
Y.Shimizu
JAXA, Japan
T.Tamura
Kanagawa University, Japan
M.Haguenauer
Ecole Polytechnique, France
W.C.Turner
LBNL, Berkeley, USA
O.Adriani, L.Bonechi, M.Bongi, R.D’Alessandro, M.Grandi,
P.Papini, S.Ricciarini, G.Castellini INFN and Universita’ di Firenze, Italy
K.Noda, A.Tricomi
INFN and Universita’ di Catania, Italy
J.Velasco, A.Faus
A-L.Perrot
IFIC, Centro Mixto CSIC-UVEG, Spain
Massimo Bongi – ICATPP – 3rd October 2011 – Como
CERN, Switzerland
4
LHCf experimental set-up
Protons
CMS
Charged particles (+)
Neutral particles
LHCf
TAN
Beam pipe
Charged particles (-)
ALICE
ATLAS
LHCb
ATLAS
140m
96mm
LHCf Detector
(Arm1)
Massimo Bongi – ICATPP – 3rd October 2011 – Como
5
Arm1 detector
• Sampling e.m. calorimeters:
each detector has two
Scintillating Fibers + MAPMT:
4 pairs of layers (at 6, 10, 30, 42 X0),
tracking measurements (resolution < 200 μm)
calorimeter towers
which allow to reconstruct 0
40mm
• Front counters:
thin plastic scintillators, 80x80 mm2
 monitor beam condition
20mm
 estimate luminosity
 reject background due to beam - residual
gas collisions by coincidence analysis
Absorber: 22 tungsten
Plastic Scintillator: 16 layers, 3 mm thick,
layers, 44 X0, 1.55 
trigger and energy profile measurement
Massimo Bongi – ICATPP – 3rd October 2011 – Como
6
Arm2 detector
• Sampling e.m. calorimeters:
each detector has two
Silicon Microstrip:
4 pairs of layers (at 6, 12, 30, 42 X0),
tracking measurements (resolution ~ 40 μm)
calorimeter towers
which allow to reconstruct 0
32mm
• Front counters:
thin plastic scintillators, 80x80 mm2
 monitor beam condition
25mm
 estimate luminosity
 reject background due to beam - residual
gas collisions by coincidence analysis
Absorber: 22 tungsten
Plastic Scintillator: 16 layers, 3 mm thick,
layers, 44 X0, 1.55 
trigger and energy profile measurement
Massimo Bongi – ICATPP – 3rd October 2011 – Como
7
ATLAS & LHCf
Massimo Bongi – ICATPP – 3rd October 2011 – Como
8
Arm1 detector
Arm2 detector
Massimo Bongi – ICATPP – 3rd October 2011 – Como
9
What LHCf can measure
Energy spectra and transverse momentum
distribution of:
 gamma rays
(E>100 GeV,
dE/E<5%)
 neutral hadrons (E>few 100 GeV, dE/E~30%)
 π0
(E>600 GeV,
dE/E<3%)
in the pseudo-rapidity range η>8.4
Multiplicity @ 14TeV
Front view of calorimeters,
@100μrad crossing angle
η
Projected edge of beam pipe
8.5
Energy Flux @ 14TeV
∞
mm
Low multiplicity
High energy flux
(simulated by DPMJET3)
Massimo Bongi – ICATPP – 3rd October 2011 – Como
10
Event categories
leading baryon
(neutron)
LHCf calorimeters
hadron event
multi meson production
π0
photon
π0 event
photon event
Massimo Bongi – ICATPP – 3rd October 2011 – Como
11
Summary of operations in 2009 and 2010
With stable beams at 450GeV+450GeV
Total of 42 hours for physics (6th–15th Dec. 2009, 2nd-3rd,27thMay 2010)
~ 105 showers events in Arm1+Arm2
With stable beams at 3.5TeV+3.5TeV
Total of 150 hours for physics (30th Mar.-19th Jul. 2010)
Different vertical positions to increase the accessible kinematical range
Runs with or without beam crossing angle
~ 4·108 shower events in Arm1+Arm2
~ 106 0 events in Arm1+Arm2
Status
Completed program for 450GeV+450GeV and 3.5TeV+3.5TeV
Removed detectors from tunnel in July 2010 (luminosity >1030 cm-2s-1)
Post-calibration beam test in October 2010
Upgrade to more rad-hard detectors to operate at 7TeV+7TeV in 2014
Massimo Bongi – ICATPP – 3rd October 2011 – Como
12
Photon energy spectra analysis
EXPERIMENTAL DATA
• p-p collisions at √s=7 TeV, no crossing angle (Fill# 1104, 15th May 2010 17:45-21:23)
• Luminosity: (6.3÷6.5) x 1028 cm-2s-1 (3 crossing bunches)
• Negligible pile-up (~0.2%)
• DAQ Live Time: 85.7% (Arm1), 67.0% (Arm2)
• Integrated luminosity: 0.68 nb-1 (Arm1), 0.53 nb-1 (Arm2)
MONTE CARLO DATA
• 107 inelastic p-p collisions at √s=7 TeV simulated by several MC codes:
DPMJET 3.04, QGSJET II-03, SYBILL 2.1, EPOS 1.99, PYTHIA 8.145
• Propagation of collision products in the beam pipe and detector response simulated by
EPICS/COSMOS
ANALYSIS PROCEDURE
1.
Energy Reconstruction: total energy deposition in a tower (corrections for light
yield, shower leakage, energy calibration, etc.)
2.
Rejection of multi-hit events: transverse energy deposit
3.
Particle identification (PID): longitudinal development of the shower
4.
Selection of two pseudo-rapidity regions: 8.81 < η < 8.99 and η > 10.94
5.
Combine spectra of Arm1 and Arm2 and compare with MC expectations
Massimo Bongi – ICATPP – 3rd October 2011 – Como
13
1 TeV π0 candidate event
scintillator layers – longitudinal development
25mm
tower
600 GeV
photon
32mm
tower
420 GeV
photon
silicon layers – transverse energy
X view
Y view
Massimo Bongi – ICATPP – 3rd October 2011 – Como
 Energy
reconstruction
 PID
 π0 mass
reconstruction
 Hit position
 Multi-hit
identification
14
Energy reconstruction

Energy reconstruction: Ephoton = f( Σ Ei ) (i = layer index)
(Ei = A x Qi determined at SPS; f() determined by MC and checked at SPS)

Impact position from lateral distribution

Position dependent corrections:
•
•
•
light collection non-uniformity
shower leakage-out (and 2 mm edge cut)
shower leakage-in
Light collection non-uniformity
Shower leakage-out
Shower leakage-in
32mm
tower
correction
2 mm
Massimo Bongi – ICATPP – 3rd October 2011 – Como
25mm
tower
15
Multi-hit rejection
 Rejection of multi-hit events is mandatory especially
at high energy (> 2.5 TeV)
 Multi-hit events are identified thanks to position
sensitive layers in Arm1 (SciFi) and Arm2 (Si-mstrip)
Arm2
Small tower
Multi-hit
detection
efficiency
Single-hit
detection
efficiency
Massimo Bongi – ICATPP – 3rd October 2011 – Como
Large tower
Arm1
Arm2
16
Particle identification
 L90%: longitudinal position containing 90% of the shower energy
 Photon selection based on L90% cut
500 GeV < EREC < 1 TeV
 Energy dependent threshold in order to
keep constant efficiency εPID = 90%
 Purity P = Nphot/(Nphot+Nhad) estimated by
comparison with MC
 Event number in each bin corrected by P/εPID
photon
hadron
 MC photon and hadron events are
44 X0
1.55 λ
independently normalized to data
 Comparison done in each energy bin
 LPM effects are switched on
Massimo Bongi – ICATPP – 3rd October 2011 – Como
17
π0 mass reconstruction
 Energy scale can be checked by π0 identification
 Mass shift observed both in Arm1 (+7.8%) and Arm2
(+3.7%)
 Many checks have been done to understand the energy
scale difference: the estimated systematic uncertainty
on π0 reconstruction is 4.2%
 Conservative approach: no correction is applied to the
energy scale, but an asymmetric systematic error is
assigned
1(E1)
R
m ≈ θ√(E1xE2)
Arm2
MC
Peak :
135.0 ± 0.2
MeV
R
140 m

140 m
2(E2)

I.P.1
Massimo Bongi – ICATPP – 3rd October 2011 – Como
18
Comparison between the two detectors
 We define two common pseudo-rapidity and azimuthal regions for the
two detectors: 8.81 < η < 8.99, Δφ = 20˚ (large tower)
η > 10.94,
Δφ = 360˚ (small tower)
R1 = 5 mm
R2-1 = 35 mm
R2-2 = 42 mm
 Normalized by the number of inelastic collisions
(assuming σine = 71.5 mb)
 General agreement between the two detectors
(deviation in small tower within the error)
Red points: Arm1 detector Blue points: Arm2 detector
Massimo
– ICATPP – 3rd
October 2011 uncertainties
– Como
Filled
area:Bongi
uncorrelated
systematic
Δφ
19
Combined photon spectra
error bars: statistical error
gray hatch: systematic error
Massimo Bongi – ICATPP – 3rd October 2011 – Como
20
Comparison with MC
magenta hatch: MC statistical error
gray hatch: systematic error
DPMJET 3.04 QGSJET
SYBILL
2.1 2011
EPOS
1.99 PYTHIA 8.145
Massimo II-03
Bongi – ICATPP
– 3rd October
– Como
21
Preliminary π0 spectra
 Mass range selection: +/- 10 MeV around
the measured peak
 Acceptance depends on energy and
transverse momentum (lowest energy is
limited by maximum angle between photons)
 Comparison with MC is on-going
 Extend the analysis
to events with two
photons in the same
tower
m ≈ θ√(E1xE2)
Eπ = E1+E2
Massimo Bongi – ICATPP – 3rd October 2011 – Como
22
Summary and outlook
 Single photon analysis at 3.5 TeV + 3.5 TeV:
• first comparison of various hadronic interaction models with
experimental data in a challenging phase space region
• very safe estimation of systematics
• no model perfectly reproduces LHCf data, especially at high energy
new input data for model developers
implications for HE CR physics under study
 Neutral pion analysis at 3.5 TeV + 3.5 TeV is in progress:
• compare pion spectra with MC
• include events with two gammas hitting the same tower
• the same analysis can be extended to η and K0 particles
 Other analysis: complete the analysis at 450 GeV + 450 GeV, neutrons,
transverse momentum distributions, extend pseudo-rapidity range,…
 We are upgrading the detectors to improve their radiation hardness
(GSO scintillators):
• we will come back on the LHC beam for the 7 TeV + 7 TeV runs
• discussion is under way to come back for possible p-Pb runs in 2013
Massimo Bongi – ICATPP – 3rd October 2011 – Como
23
Backup
Open Issues on UHECR spectrum
AGASA Systematics
Total
±18%
Hadr Model
~10%
(Takeda et al., 2003)
M Nagano
New Journal of Physics 11 (2009) 065012
Depth of the max of
the shower Xmax in
the atmosphere
AUGER
HiRes
Massimo Bongi – ICATPP – 3rd October 2011 – Como
25
Front counters
• Thin scintillators with 8x8cm2 acceptance,
which have been installed in front of each main detector.
Schematic view of
Front counter
• To monitor beam condition.
• For background rejection of
beam-residual gas collisions
by coincidence analysis
Massimo Bongi – ICATPP – 3rd October 2011 – Como
26
Detector vertical position and acceptance
Remotely changed by a manipulator( with accuracy of 50 mm)
Viewed from IP
G
Distance from
neutral center
Data taking mode
with different position
to cover PT gap
Beam pipe aperture
N
L
Neutral flux center
All  from IP
7TeV collisions
L
Collisions with a crossing angle
lower the neutral flux center thus
enlarging Pt acceptance
N
27
Expected results @ 14 TeV collisions
Energy spectra and transverse momentum
distribution of:

• photons
(E > 100 GeV): E/E < 5%
• neutral pions (E > 500 GeV): E/E < 3%
• neutrons (E > few 100 GeV): E/E ~ 30%
in the pseudo-rapidity range  > 8.4
n
0
Massimo Bongi – ICATPP – 3rd October 2011 – Como
28
LHCf energy resolution
2.5 x 2.5 cm2 tower
2.0 x 2.0 cm2 tower
Energy resolution < 5% at high energy,
even for the smallest tower
Massimo Bongi – ICATPP – 3rd October 2011 – Como
29
Arm1 position resolution
200 GeV electrons
σX[mm]
Number of event
σX=172µm
x-pos[mm]
E[GeV]
Number of event
σY[mm]
σY=159µm
y-pos[mm]
E[GeV]
Massimo Bongi – ICATPP – 3rd October 2011 – Como
30
Arm2 position resolution
200 GeV electrons
Position Resolution X Side
120
Data
Resolution (microns)
σX=40µm
σX[µm]
100
Simulation Spread Out
80
60
40
20
0
0
x-pos[mm]
50
100
E[GeV]
150
200
250
Energy (GeV)
Position Resolution Y Side
160
140
Data
Simulation Spread Out
Resolution (microns)
σY=64µm
σY[µm]
120
100
80
60
40
y-pos[mm]
20
0
Alignment has been taken
into account
Massimo Bongi – ICATPP – 3rd October 2011 – Como E[GeV]
0
50
100
150
Energy (GeV)
200
250
31
Estimation of pile-up
•When the circulated bunch is 1x1, the probability of N collisions per crossing is:
N e  
P( N ) 
N!
L× s
l=
f rev
•The ratio of the pile-up event is:
P(N ³ 2) 1- (1+ l )e- l
Rpileup =
=
-l
P(N ³1)
1- e
•The maximum luminosity per bunch during runs used for the analysis is 2.3x1028cm-2s-1
•So the probability of pile-up is estimated to be 7.2%, with σ=71.5mb and frev = 11.2 kHz
•Taking into account our calorimeter acceptance for an inelastic collision (~0.03) only 0.2%
of events have multi-hit due to pile-up
Massimo Bongi – ICATPP – 3rd October 2011 – Como
32
Luminosity estimation
• Luminosity for the analysis is calculated from Front
Counter rates:
L = CF ´ RFC
•The conversion factor CF is estimated from luminosity
measured during Van der Meer scan
VDM scan
LVDM = n b f rev
I1I2
2ps xs y
Beam sizes sx and sy measured
directly by LHCf
Massimo Bongi – ICATPP – 3rd October 2011 – Como
33
0 mass vs 0 energy
Arm2 data
No strong energy
dependence of
reconstructed mass
Massimo Bongi – ICATPP – 3rd October 2011 – Como
34
2 invariant mass spectrum @ 7 TeV
Arm2 detector, all runs with zero crossing angle
True η mass:
547.9 MeV
MC reconstructed η mass peak:
548.5 ± 1.0 MeV
Data reconstructed η mass peak: 562.2 ± 1.8 MeV (2.6% shift)
Massimo Bongi – ICATPP – 3rd October 2011 – Como
35
Effect of mass shift
 Energy rescaling NOT applied but included in energy
error
 Minv = θ √(E1 x E2)
– (ΔE/E)calib = 3.5%
– Δθ/θ = 1%
– (ΔE/E)leak-in = 2%
=> ΔM/M = 4.2% ; not sufficient for Arm1 (+7.8%)
±3.5% Gaussian probability
±7.8% flat probability
135MeV
Quadratic sum of two errors is
given as energy error
145.8MeV
(to allow both 135MeV and
(Arm1 observed) observed mass peak)
Massimo Bongi – ICATPP – 3rd October 2011 – Como
36
Systematic uncertainties
• Energy reconstruction (detector response, from beam test at SPS): 3.5%
• Multi-hit rejection (estimated by comparing true MC and reconstructed MC spectra
after MH cut): from 1% for E < 1.5 TeV to 20% at E = 3 TeV
• PID (by comparing 2 different approaches): 5% for E < 1.7 TeV, 20% for E > 1.7 TeV
• Beam center position (by comparing position measured by LHCf and by Beam Position
Monitors): 5÷20 % varying with energy and pseudo-rapidity
• Luminosity (Front Counter measurement during Van der Meer scans): 6.1%, it causes
an energy independent shift of spectra, not included in photon spectra
• Energy shift (π0 mass shift, asymmetric): 7.8% Arm1, 3.7% Arm2
Total energy scale systematic:
-9.8% / +1.8% for Arm1
-6.6% / +2.2% for Arm2
Systematic uncertainty on spectra is estimated from the difference
between normal spectra and energy scaled spectra
Massimo Bongi – ICATPP – 3rd October 2011 – Como
37
Background
1. Pile-up of collisions in one beam crossing
 Low Luminosity fill, L=2.3x1028cm-2s-1
 7.2% pile-up at collisions, 0.2% at the detectors.
2. Collisions between secondary's and beam pipes
 Very low energy particles reach the detector (few % at 100GeV)
3. Collisions between beams and residual gas
 Estimated from data with non-crossing bunches.
 ~0.1%
Secondary-beam pipe backgrounds
Beam-Gas backgrounds
Massimo Bongi – ICATPP – 3rd October 2011 – Como
38
Energy spectra at 900 GeV
gamma-ray like
hadron like
Arm1
Arm2
Acceptance is different for the two arms.
rd Octoberlike
Spectra are normalized by
# ofBongi
-ray
and– 3hadron
Massimo
– ICATPP
2011 –events.
Como
Only statistical
errors39
are
shown
Radiation damage studies
 Dose evaluation on the basis
 test of Scintillating fibers and scintillators
of LHC reports on radiation
environment at IP1
 ~ 100 Gy/day @ 1030 cm-2s-1
luminosity are expected
 ~ 10 kGy during few months
operation lead to ~ 50% light
output decrease
 continuous laser calibration
to monitor scintillators and
30 kGy
Massimo Bongi – ICATPP –
3rd
correct for the decrease of
light output
October 2011 – Como
40