What have we seen so far ? Position and momentum measurement - little or no loss of energy Time measurement - little or no loss of energy Energy measurement - loss of energy Today ‘s goals : Identification • general principles • muon/electron/pion/jets • b or c quark • Cherenkov radiation • Transition radiation Literature: -paper linked in ISIS page by the slides -D.Green pg 55-87
30
Embed
What have we seen so farxella/lecture_4Mar2009.pdf · 2009. 3. 6. · Best vertex resolution so far achieved -> best tagging of b and c quark originated jets. First, construct f i(r)
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
What have we seen so far ?
Position and momentum measurement - little or no loss of energyTime measurement - little or no loss of energyEnergy measurement - loss of energy
Today ‘s goals :
Identification• general principles• muon/electron/pion/jets • b or c quark• Cherenkov radiation • Transition radiation
Literature:-paper linked in ISIS page by the slides -D.Green pg 55-87
What do we need to measure the mass, henceidentify a particle uniquely ?
E p
Ideally: measure E and p very precisely , then calculate m(for particles decaying, sum up all E_i and p_i)
So we do not need any particle identification
How can we identify a particle ?
Energy(1)
Energy(2)
Momentum(2)
Momentum(1)
Example : ALEPH experiment at LEP
The goal of the ALEPH experiment is to proof thate+e- → Z0 → ff exists, and ff is the decay of a particle, Z0 boson
Ok, I clearly see 2 objects, so I should do (E(1)+E(2))^2 - (p(1)+p(2)) ^2 = M^2
But :
E(i) << p(i) , so ???
I need to understand a bit more what I see
So if I doE(1)=sqrt( p(1)^2 + m_muon ^2) and I then reconstruct M, what doI get ?
if 1, 2 are muons then I know they are M.I.P.and E in the calorimeter is not full E of muon , and I should trusttracker for full P.This interpretation is also confirmed by two muonsleaving signal in the outer chambers
E(1)<<p(1), E(2)<<p(2) !
What should I make of this ?
Resonance shape ! It’s a Z0 we observed !
They also help with identification, not all detectors have an ad-hoc particle identificationdetector or rely only on that one for identification
ALEPH didn’t have one ad-hoc detector for identification only
• Energy(1)/momentum(1) < 1.
• Energy(1)/momentum(1) ~ 1.
Energy(1)
Momentum(1)
Energy(1)
Momentum(1)
• Stops in electromagnetic calorimeter
• Does not stop entirely in electromagnetic (EM) calorimeter• There is signal around in EM and has no track attached
electron
pion
photons
τ -> e νντ -> π π0 ν
(what do we expect the total energy to be ?)
electron
positron
20 50 …
Calorimeter helps:
E.g. e/π-Lateral Shape-E/p
Tracking helps:
Identification is not simply done “by eye”, but combines severalinformations from several parts of the detector
• We have seen the case of the muon (E not fully collected)• electron, pion (again, we use ALEPH example) ?
σ(pT)/pT = 0.6 10-3 pT
calorimeter Aleph tracking
Try eg. p=p_T= 45 GeV
Using measured E and p is actually worse than using best of the twoand identification (-> mass well known) -> energy flow
Identification methods :
Use the fact that dE/dx depends on velocity, hence particle mass (fora given momentum)
Use the fact that in 20 cm of iron the probability for a electron to stop is much higher than for a pion (brehmstrahlung goes like 1/mass 2)so energy collected and size of shower is quite different
Exercise :
Try to identify the following events from ALEPH detector
Remember: physics process happening is
e+ e- → Z0 / gamma → f f with any f
3 minutes for each figure, discuss 2 and 2 andwrite down your identification
Discussion over evet displays.
My identification is :1) electron positron2) tau tau3) muon muon4) q q5) tau tau6) Mu mu gamma
For Alpeh performance, see paper linked from ISIS page for this week’s lectures
How to find a secondary vertex (http://www.phys.ufl.edu/~avery/fitting.html)
B and c flavour tagging: identifying long lived quarks
2mm
SLD detector
ee→Z0→bb
SLD : e+ and e-linear collider, c.m.s. 91 GeVZ boson production, similar program to LEP experiments
Best vertex resolution so far achieved -> best tagging of b and c quarkoriginated jets.
First, construct fi(r) = probability tube in 3-D for the i-th track trajectory
In ee collisions at 91 GeV , b and c quark from Z0 live relatively long (b ~1 mm)
Track Functions Vertex Functiontracks tracks
IPB
3-D Spatially resolved clusters of V(r) maxima form candidate vertices.
No combinatorics, you just see the right pattern
Third, tracks are attached to these resolved regions to form a set oftopological vertices
Need carefull tuning ofthese T and L/D cutsbefore using them for agiven vertex detectordesign
All non-primary vertices foundmust be n-prong with n ≥ 2
BUT
For IP_B_D decay chain canhave 1-prong B or D decay
Can consider n-prong ZVTOPvertex as a seed vertex
A track not directlyassociated with the Primary
or Seed vertex, but withT<Tmax and L/D>L/Dmin, is
likely to come from B decaychain
Finally variables coming out for tagging are
Apply a kinematiccorrection to MVTX to
partially recover effectof missing neutral
particles:
PTmiss
MPT
flavour tagging informations available are :
Vertex MassVertex MomentumDecay LengthDecay Length significanceN tracks associated to sec. Vtxs …..
Finally, Vertex Mass reconstructed
SLD Ghost track Algorithm
Angle between true B flight and jet axisAngle between true B flight and ghost track
Will improve these mostly
D
B
• • • •
In e+e- →Z0/γ→qq events, at 91 GeV c.m.s. energy (simulated data)
SLD-c
SLD-b
Efficiency of determing flavour “b” for a jet
Efficiency of determing flavour “c” for a jet
Results from SLD CCD-based vertex detector (stars) and from a new designfor a CCD-based vertex detector with parallel column readout
50 µs (8 msNLC)
216 msreadout time (1 layer)
0.96 (3 hits)0.90 (2hits)
cos(θ) max0.10.4layer thickness (%Xo)1528inner layer radius(mm)53n. of layers799307n. of pixels (106)27.512.8CCD active area (cm2)12096CCDsFuture LCSLDDetector