Particle Identification: Computer reconstruction of a UA1 event with an identified electron as a candidate for a W >eν event

Particle Identification: Computer reconstruction of a UA1 event with an identified electron as a candidate for a W >eν event

Valuable particles at hadron colliders are the electron e ± for W ±! e ± & Z 0! e + e and the muon µ ± for W ±! µ ± & Z 0! µ + µ All high-energy collider detectors focus on these because almost everything (Technicolor, SUSY, W, Z,...) eventually decay down to W sandzs. Next, sforthecleanesthiggsdecay,h 0!. although single sareincreasinglydi cult to separate from 0! due to the small 0 mass, m =0.135 GeV/c 2 : m/e 0.01 at E 15 GeV.

Telling an e ± from a ± by depth development Electromagnetic showers develop in depth over 10-20 X 0 (radiation lengths) while hadrons develop in depth over 5-10 int (nuclear interaction lengths).

Longitudinal (depth) fluctuations: compare every other layer

Telling an e ± from a ± by a pre-shower detector apre-showerdetectorusesonlytheverybeginning of a shower, since an electron will immediately begin to bremsstrahlung in the Pb generating a large number of low-energy s, and losing on the average about 90% of its energy in 2 X 0. int X 0 0.35 Z. Therefore, U and Pb are best for e separation with int /X 0 30.

Telling an e ± from a ± by lateral development The transverse development of electromagnetic showers, like the depth development, is both narrower and similar shower-to-shower, and these properties can be exploited in any calorimeter that is laterally segmented on a scale of 2-4 X 0.

Telling an e ± from a ± by E/p The e ± and ± momenta will be measured in the tracking system and their energies measured in the electromagnetic calorimeter (first 20 X 0,or 1 int ).

E/p - E70 Fermilab

Telling an e ± from a ± by dual-readout, Svs.C

Telling an e ± from a ± by fluctuations in (S k C k ) The Svs.Cplot of Fig.?? is for the whole shower, where S and C are the sums of the energies in all the channels that are believed to be activated by the particle, S = NX S k and C = NX C k. k=1 k=1 If the shower is electromagnetic, then S k C k for all channels k. Achi-squared statistic is constructed that tests this: 2 = NX S k C k! 2, k=1 k where k is the expected rms variation of (S k C k )forelectrons, 2 k 0.1(S k + C k ).

Chi-squared for (S-C)

Telling an e ± from a ± by the time development of the scintillation signal Most of these discriminators are independent, and therefore their rejection factors multiple.

Identifying a! an e without a charged track pointing to it. PS shows e - di erences. ( + )vs.e

Telling a µ ± from a ± Muon identification and measurement in an iron (Fe) absorber A solenoid B field must be returned to the other end, and an iron yoke is almost always used as the high-permeability flux channel. This iron volume is simultaneously useful for mechanical structure of the whole detector and, finally, as achargedpion( ± )absorberthatallows,mostly,onlymuons(µ ± )topenetrate the 2-3 meters of iron. The e cacy of this filter depends somewhat on the energies of the pions, but is mostly limited by the punch through of actual hadronic pions and protons which can make repeated elastic or di ractive scatters in the iron. It should be noted that pions develop hadronic showers consisting partly of many more lower-energy pions and kaons that decaytomuons, ±! µ ± µ and K ±! µ ± µ,andthereforetheironyokeservestoincreasethe punch-through of low energy muons. The measured momentum of the muon in the iron yoke, however, is fundamentally limited to 10%, or worse, independent of the precision of the tracking spatial measurements, independent of the momentum, and insensitive even to the depth of iron over which the track measurements are made, as shown in the following calculation:

Only µ + and µ - left out here.

Telling a µ ± from a ± The variation of the µ ± momentum with the sagitta is p = p 2 8 [ ] s, (1) 0.3B`2 and in this multiple-scattering dominated iron volume, the sagitta variation has two contributions, the usual one from the chamber spatial precision ( s x )and the second from multiple scattering given by s rms,so s = s x + s MS, and where r 0.0136 (GeV/c) ` s ms = s rms = 4 p `. (2) 3 p X 0 X 0 is the radiation length of the medium (in this case, 17.6 cm for Fe)and ` is the depth of the Fe absorber. The µ ± is very energetic so 1, and a quick estimate of the magnitudes of theseq two contributions to s, forµ ± at any energy, yields s x s MS,sothat p = ( p)and 2 p = p 2 " 8 0.3B`2 # " r 0.0136(GeV/c) ` 4 p ` 3p X 0 #. (3) The radiation length of Fe is X Fe 0 0.176 m and the maximum magnetic field is at saturation in iron, B 1.8T, so that p p ' p`(m) 0.15 (in an iron spectrometer), (4) asimpleresultindependentofmomentumandonlyweaklydependent on the

Neutrino experiments in large detector volumes track muons: WA1

Telling a µ ± from a ± Muon identification and measurement in an iron-free detector Only one major detector is iron-free, the ATLAS experiment at the LHC at CERN, with its hadronic calorimeter as the pion filter. It will be of great interest to compare the muon physics from the CMS and ATLAS experiments inthe coming years. Muon ID by momentum balance In a dual-solenoid iron-free detector, the momentum of a µ ± from the IP will be measured in the tracking system (p µ1 ), any energy loss (by ionization or radiation) will be measured inthecalorimeter ( E µ ), and its momentum measured again (p µ2 )intheannulusbetweenthe solenoids. A muon candidate must satisfy momentum balance, p µ1 E µ +p µ2, where the precision of this balance depends on the momentum of themuon,and the radiation of the muon inside the solenoid and its cryostat, and is typically 3 5%. This balance can reject pions against muons to about a factor of 30 for decay muons from W and Z 0.

Outer solenoid Inner solenoid Muon tracking volume Dual readout Calorimeters 19

New magnetic field, new ``wall of coils, iron-free: many benefits to muon detection and MDI, Alexander Mikhailichenko design

Muon trajectories from the interaction point B~-1.5 T B=3.5 T

< B > T 4th Concept Muon Tracking Field Dual solenoid tracking along muon trajectories in the annulus between solenoids. < B d > T m cluster counting drift tubes for muon tracking.

Telling a µ ± from a ± Muon identification and measurement in an iron-free detector Muon ID by dual-readout of Cerenkov and scintillation light The separation of µ ± radiation and ionization energy losses in a dense medium has a consequence for µ ± identification and discrimination against ±.thedi erence of the scintillation (S) and(c) signalsareplottedagainsttheiraveragefor80 GeV particles. The µ ± s are well separated from the ± s, and this remains so from the lowest measured energies (20 GeV) to the highest (200 GeV) Since the important µ ± are usually isolated this method can be e ective, whereas in a dense jet environment the calorimeter identification of the S and C signals will be di cult. These two methods, muon momentum balance and (S-C), are independent and their pion rejection factors multiplied.

The Cerenkov signal from an approximately aligned, nonradiating muon is zero Photon at Cerenkov angle Cerenkov fiber Muon Numerical aperture of fiber: capture cone All of the Cerenkov light of an approximately aligned muon falls outside of the numerical aperture.

S+C (GeV) Muons (40 GeV) & Pions (20 GeV) µ S-C (GeV) S-C (GeV)

Muons and Pions (80 GeV) µ S-C (GeV) S-C (GeV)

Muons and Pions (200 GeV) π µ S-C (GeV) S-C (GeV)

Muons and Pions (300 GeV) µ π S-C (GeV) S-C (GeV)

Four 5-GeV muons through detector as test Muons are clean and obvious; Acceptance at 5 GeV is good; Momentum and energy measurements must add up for a real muon; GEANT simulation in very good shape in a very short time; Still, there is more fun work to do.

Low-momentum (0-10 GeV/c) charged hadrons: ±,K ±,p, p de/dx: Specific ionization / 1/ 2 Cerenkov angle: cos C =1/n Time-of-Flight (ToF): ct = L/ =(L/p) p m 2 + p 2 c[t 2 t 1 ] (L/2p 2 )[m 2 2 m 2 1] All spend on velocity: = p/e = p/ p m 2 + p 2

de/dx measured on proportional wires in 8.6 atm. Ar gas TPC e, µ, and deuteron, also

Cerenkov: good for low momentum tracks, much harder for high momentum. Good optics problem:

KLOE at Frascati, a truly beautiful detector

γ vs. π 0 > γγ

n, K 0 L? Probably not; n, K 0 L should behave just like p, K+ in a calorimeter, but too close to call. We will test pvs.pi +.

u, d, s quarks from g gluons. At a hadron collider, the overwhelming dominant jet background is gluons from gg! gg Physics particles are almost always quarks from W! qq, Z! qq This the most important problem for physics at a hadron collider: Di erences: gluons! higher multiplicity of low energy particles. W sandzs: low energy (40-45 GeV in center-of-mass) quark jets. Somebody someday has to find a good solution to this problem... and define the detector that will accomplish it.

W ±! q q! jet+jet and Z 0! q q! jet+jet Measure jets with /E 30%/ p E

Identifying weak s-baryon decays: (sdu), (sqq), (ssq), and (sss)! decays lengths of c of many centimeters. Identifying weak heavy quark decays: B(b q) andd(c q)! decays lengths of c millimeters 10 4 3 10 10 2 b-jets 10 c-jets 1 light-jets -1 0 1 2 3 4 5 6 7 8 log(1+d/!)

Identifying the ± lepton: Critical decay for polarization in H 0! + decay is!! 0! (three objects in the detector)

This has been a fast look at a vast array of particle identification techniques. We have left out Transition Radiation, have not pursued the very interesting aspects of Cerenkov light for particle ID, and have not pointed out what are the limitations of each, e.g., what defeats each particle ID scheme.