LHCf StatusOscar Adriani

CSN1,MIlano, 26 Marzo 2013

Summary of operations in end 2012- beginning 2013 (after the Torino meeting)

Arm2 has been re-installed in the TAN region in December 2012

January-February 2013: p/Pb run Arm2 will be removed from the TAN in April 2013 The Arm2 upgrade for the 13 TeV run will be done

in Florence in 2013 in strict collaboration with Japanese colleagues

Some more details will be given in the next slides

Re-Installation issues Arm2 has been successfully re-installed in the TAN

during the technical stop foreseen at the end of the p/p run December 18th

We have modified the LHCf support structure and cabling to significantly reduce the installation required time

Mechanical survey has been done in 2 steps: Internal LHCf survey on ground LHCf survey wrt to LHC: done on December 18th in the

TAN area No big problem of radiation, the installation was

completely safe (Thanks to Raffaello and Sako!!!)

Discussions and agreements with ATLAS (I)

ATLAS trigger can not be sent to LHCf due to timing problem

LHCf Level1 Trigger signal has been sent to ATLAS for the whole p/Pb running period

ATLAS has properly prescaled the LHCf trigger signal Prescaling factor depend on the running conditions

LHCf has recorded in the data stream all the counters and has used all the signals necessary to off-line identify the common events Event Counter Reset Atlas L1ID Bunch ID

Discussions and agreements with ATLAS (II)

Proton remnant side – Invariant cross section for isolated g-raysUsing only the LHCf informations

What happens if know the Impact Parameter?

Ideal case, assuming that we can precisely know the Impact Parameter b (in fm) on event by event basis

What happens if know the Impact Parameter?

What happens if know the Impact Parameter?

What happens if know the Impact Parameter?

What happens if know the Impact Parameter?

Combination of different impact parameter bins

In real life b should be estim

ated by

using the Atlas information (ce

ntrality)

The difference between models is

enhanced by the knowledge of the

impact parameter

What happens with LHCf on Pb remnant side?

Nominal vertical position (Dy=0 cm)

Shifting up by Dy= +2.5 cm

The small calorimeter tower remain in the region not screened by the narrow elliptical shape of the beam pipe at D1 magnet We can take good data with reasonable number of hits!

LHCf operation in p – Pb runs at √sNN= 5 TeV

Pb p

IP8IP2 IP1Arm2

p Pb

IP8IP2 IP1Arm2

Proton remnant side

Lead remnant side

LHCf operation in p – Pb runs at √sNN= 5 TeV

#Eve

nts (

Mill

ions

)

p-remnant side

Pb-remnant side

Beam reversal

20 Jan 27 Jan. 01 Feb.

200 Millions triggered events!!!!

Summary of LHCf p-Pb runs L = 0.5x1029 – 1x1029cm-2s-1

b* =0.8m, 145mrad crossing angle Not good for LHCf…. We didn’t succeed to get a dedicated high b* run due to

the lack of time 338p+338Pb bunches (min.DT=200ns), 296

colliding at IP1 10-20kHz trig rate downscaled to ~700Hz 20-40Hz ATLAS common trig

Coincidence operation was successful!!! Data both at p-side (20Jan-1Feb) and Pb-side (1fill, 4Feb)

Operation at Pb-remnant side

A high multiplicity event (Pb-side)

p Pb

IP8IP2 IP1Arm2

MC (Pb-remnant)

3.5cm, 4.0cm

Proton-Proton Collision at √s = 2.76 TeV

We also profited of the ‘calibration’ run at √s = 2.76 TeV that has been done following the ATLAS and CMS requests

4 hours operation on 14 Feb. 2013 successfully done.

These data will allow a better study of the energy Scaling by comparing different c.m. energy (0.9 TeV, 2.76 TeV, 7 TeV, 13 TeV)

Data list of LHCf

p-p, √s=900GeV, 2010

✔ (event flags)

p-p, √s=2.76TeV, 2013

✔ LHCf triggers

p-p, √s=7TeV, 2010 ✔ ✔ (event flags)

p-p, √s=13 TeV, (2015)

✔ ✔ LHCf triggers

p-N,O, (>2019) ✔ ✔ LHCf triggers

p-Pb, √sNN=5TeV, 2013

✔ ✔ LHCf triggers

p-p 400GeV, p-Aat RHICH (???)

✔ ✔ PHENIX, STAR

γ, n π0With ATLAS

Orange:Future operations

Black:completed operations

xF = E/E0

Playing a game with air shower development:effect of forward meson spectra

• DPMJET3 always overpredicts production• Filtering DPMJET3 mesons

• according to an empirical probability function, divide mesons into two with keeping pT

• Fraction of mesons escape out of LHCf acceptance

• This process• Holds cross section• Holds elasticity/inelasticity• Holds energy conservation• Changes multiplicity• Does not conserve charge event-by-event

E=E1+E2

E1E

2

xF = E/E0

pT

An example of filtering

π0 spectrumphoton spectrum

DPMJET3+filter2.5x1016 eV proton

~30g/cm2

Apart from this ‘game’ we are in strict contacts with model developers to help them improving their codes.Few dedicated workshops have been organized to put theorists and experimentalists in contact

π0 spectrum and air shower

Vertical Depth (g/cm2)

AUGER, ICRC 2011

100 g/cm2

30 g/cm2

Other analyses and future activities….

Joint analysis with ATLAS … data ready 14 TeV p-p in 2015 … detector upgrade on going Neutron spectra in 7TeV p-p … analysis on going Light nuclei at LHC, RHIC??? … possibility in discussion

LHCf preparation for the 14 TeV p-p run

Calorimeter radiation hardening by replacing plastic scintillator with GSO Production and laboratory tests of the new scintillators in Japan is

finished for Arm1 and in progress for Arm2 Beam test at Ion facility (HIMAC) for Arm1 has been done in June

2012 Arm1 has been re-assembled in Florence starting from end of

June 2012 Same procedure will be followed in 2013 for the Arm2

detector Upgrade of the silicon positioning measurement system

Rearranging Silicon layers for independent precise energy measurement Increase the dynamic range to reduce saturation effects

Test Beam at LNS for the absolute energy calibration of the silicon system is being requested

Why neutron measurement is important for CR physics

Auger hybrid analysis• event-by-event MC

selection to fit FD data (top plot)

• comparison with SD data vs MC (bottom plot)

• Clear muon excess in data even for Fe primary MC

The number of muons increases with the increase of the number of baryons! => importance of direct baryon measurement

Neutron Spectra at 7 TeV pp (models)

Model predictions

Model predictions smeared taking into account the LHCf energy resolution

Life is not easy…..

1 TeV neutrons simulated with 2 different hadronic interaction models used in the detector simulation

Other possibile future runs? Possibility to use LIGHT IONS in LHC from

2016/2017? Light Ion source setup is ongoing because of SPS

interest RHIC run in 2015/2016 is under discussion… Please stand by a little bit to see how things are

evolving!!!!

Il calcolo per LHCf A settembre la CSN1 ci ha ‘suggerito’ di muoversi nella

direzione di utilizzare le risorse di calcolo del CNAF (nonostante le richieste estremamente limitate di 15 kEuro)

In questi mesi, con l’aiuto di Vincenzo Vagnoni e con il supporto di Luca Dell’Agnello, abbiamo sistemato le infrastrutture tecniche necessarie per: Generazione (per almeno 4 modelli di interazione adronica) End2End (trasporto beam-pipe) DoubleArm (simulazione del rivelatore)

Compilatori, spazio di storage, creazione degli account e delle code per I jobs, etc.

Il sistema ora e’ ‘pronto per partire’

Necessita’ Almeno 4x107 eventi

Generazione: 35 kB/event 0.1 sec/event

Trasporto: 100-500 kB/event 100-500 sec/event

Simulazione: 20 kB/event 10 sec/event

CPU: Almeno 4x109 secondi estendibili a 1.3x1010 sec se ci fosse la

necessità di avere 108 eventi per un modello Storage:

20-30 TB

Come sta andando Siamo in contatto con CNAF per finire di risolvere i

problemi tecnici rimasti La procedura e’ stata faticosa, ma alla fine siamo

(quasi) arrivati….

Non mi e’ chiaro come ora sia necessario procedere con la commissione per pagare le risorse CNAF….

Conclusions Re-installation in the tunnel and p/Pb run went very

smooth p/Pb and neutron analyses are on-going Atlas joint analysis is ready to start Arm2 upgrade will be completed in 2013 Computing system at CNAF is available Ready to take data at 14 TeV And…. Possible Light Ions runs at RHIC/LHC are under

investigation

Next week we will have the LHCf meeting in Nagoya

Spares slides

Miscellanea IV: LHCf computing Lo scorso anno abbiamo presentato un piccolo modello di calcolo per far fronte

alle esigenze di simulazione e ricostruzione di LHCf per il run p-Pb di cui siamo responsabili I referee ci hanno finanziato una parte di quello richiesto rimandando a

quest’anno la seconda parte a fronte di stime più precise per consentirci la produzione dei plot per la LOI

Il data set per la LOI è stato prodotto interamente in Italia e le tre macchine acquistate sono state fondamentali

Abbiamo fatto i primi test di simulazione completa con p-Pb 500 KB per evento e 570 sec/evento con la simulazione completa 20 KB per evento e 22 sec/evento se applichiamo dei tagli cinematici abbastanza

duri (eccessivi per quello che vorremmo fare) Una via di mezzo tra queste due, dell'ordine dei 100 KB e 100 sec/evento e' quella

piu' realistica senza perdere informazioni di fisica rilevanti. Noi abbiamo bisogno di produrre come minimo 107 eventi per ciascuno dei

modelli studiati (finora 5) Poichè le stime dello scorso anno, basate sulla sola generazione erano ben

più ottimistiche di quello che abbiamo ottenuto ora, chiediamo il completamento delle risorse. Per il disco cercheremo di utilizzare risorse presenti in sezione ma abbiamo bisogno di CPU dedicate 15 Keuro per l’acquisto delle CPU

Radiation hardness of GSO

No decrease up to 1 MGy

+20% increase over 1 kGy (τ=4.2h recovery)

2 kGy is expected for 350nb-1 @ 14TeV pp)

１ kGyNot irradiated ref. sample

Irradiated sample

τ~4.2h recovery

K. Kawade et al., JINST, 6, T09004, 2011

Dose rate=2 kGy/hour(≈1032cm-2s-1)

Proton-remnant side – photon spectrumSmall tower Big tower

Proton-remnant side – neutron spectrumSmall tower Big tower

35% ENERGY RESOLUTION IS CONSIDERED IN THESE PLOTS

What LHCf can measure in the p+Pb run (2)Study of the Nuclear Modification Factor

Nuclear Modification Factor measured at RHIC (production of p0): strong suppression for small pt at <>=4.

LHCf can extend the measurement at higher energy and for >8.4Very important for CR Physics

Phys. Rev. Lett. 97 (2006) 152302

Lead-remnant side – multiplicityPlease remind that EPOS does not consider Fermi motion and Nuclear Fragmentation

n

g

Small tower Big tower

π0 results: Data vs MC

π0 results: Data/MCSubmitted to PRD (arXiv:1205.4578).

<pT> distribution Three different approaches used to derive the average transverse momentum, ⟨pT⟩1. by fitting an empirical function

to the pT spectra in each rapidity range (exponential distribution based on a thermodynamical approach)

2. By fitting a gaussian distribution

3. by simply numerically integrating the pT spectra

Results of the three methods are in agreement and are compared with UA7 data and hadronic model predictions.

Two UA7 and LHCf experimental data show the same trend→ no evident dependence of <pT> on ECMS.

YBeam=6.5 for SPSYBeam=8.92 for7 TeV LHC

Comparison wrt MC Models at 900 GeV

small-η

= Large tower

big-η =Small tower

A jump back to g analysis: Comparison btw 900GeV and 7TeV spectra

Normalized by the number of entries in XF > 0.1 No systematic error is considered in both collision

energies.

XF spectra : 900GeV data vs. 7TeV data

Good agreement of XF spectrum shape between 900 GeV and 7 TeV. weak dependence of <pT> on ECMS

Preliminary

Data 2010 at √s=900GeV(Normalized by the number of entries in XF > 0.1)Data 2010 at √s=7TeV (η>10.94)

Coverage of the photon spectra in the plane Feynman-X vs PT

900GeV vs. 7TeVwith the same PT region

900 GeV Small+large tower

Neutron Detection Efficiency and energy linearity

Efficiency at the offline shower triggerFlat efficiency >500GeV

%

Linear fitParabolic fit

Energy and Position Resolution

X Y

Neutron incident at (X,Y) = (8.5mm, 11.5mm) ~1mm position resolutionWeak dependence on incident energy

We are trying to improve the energy resolution by looking at the ‘electromagneticity’ of the event

K0 analysis

K0 Acceptance