First look at non-Gaussian tails with the new Reconstruction Stathes Paganis

First look at non-Gaussian tails with the new Reconstruction

• Stathes Paganis

• Univ. of Sheffield

• LAr-H8 Working Group, 18-Oct-05

Data and MC: as in Gargnano

• Runs: Period 5 (20,50,100,180GeV)

• Proposed by Isabelle W. from August, no extra Al material.

• Simulation: 10.5.0

• PS E-field extends 13/11 longer in MC.
• Added 0.15X0 close and 0.15X0 far

• Recon: 10.5.0 + Tags by Marco

• OFCs : TB04-9
• Pedestals: HEAD, Ramps, ADC2uA: TB04-default
• uA2MeV: “new set” TB04-default

• Final Analysis (Gargnano):

• 0.91*Estrips (Data: first rough cross-talk approximation)
• 0.84*11/13*Eps (MC: to account for E-field extend)
• 0.97*Erec (MC: just a normalization)

100GeV runs beam profiles (reweighted)

Uncorrected energy distribution

100GeV: Energy vs eta and phi profiles

Eta vs Phi profile

100GeV: Sampling energies

Observations on the tail events (50,100GeV run)

• Cell energies: Ecell22 significantly lower

• Middle Energy: significantly lower

• Strips Energy: in very good agreement (!)

• PS energy: shifted higher

• Back energy: significantly lower

• Eta of cluster: slightly shifted towards lower eta

• Phi of cluster: slightly shifted towards lower phi

• Tile Energy: small deposition in the tile

Based on these observations:

• Cannot be pions: (observations 4 , 8)

• Cannot be lower E electrons: (4, 8)

• They look like early showers: (4, 5)

• They are not in our simulation.

• Reminder: in photon runs, electrons show no tails

• My (strong) conclusion:

• Most of the tail comes from early showers far upstream the calorimeter. This contribution is not simulated.

• To check the theory we need an early shower tagger:

• (a scintillator with a few cm hole would do; muHalo)

Ecluster vs muHalo

Tail reduced, eta dependence smaller

Middle Sampling: Before/after muHalo cut:

Conclusions/Plans

• Part of the energy tail seems to come from early showering far upstream.

• The tail has eta dependence

• Addition of material in the MC will destroy the existing good agreement (since Gargnano)

• Unless we add more material at lower eta!

• Dip structure in phiRec remains a puzzle. It looks like there is more (non-simulated) material in front of the strips in this region

• It is clear that at some point we will have to decide on certain (optimum) cleaning cuts.

Extra slides

muHalo and eta cuts

Ecell32 energy vs phi