B Physics Experiments

1
B Physics Experiments
TASI June, 2000
Sheldon Stone Syracuse University
2
Introduction Objectives

• Understand how a modern HEP B Physics experiment
  works
• Understand how the detector works
• Understand how the data is analyzed
• Understand what final states are most useful
  (i.e. easiest to deal with)
• Understand what mistakes can be made

3
Deconstructing The Detector
4
Deconstructing The Detector
1 m
Whats missing?
5
Deconstructing The Detector

• What do we want to measure?
• To look for physics results we need to find
  specific decay modes of Bs
• Sometimes we are interested in inclusive decays,
  i.e. B?Xe-n, often we are interested in exclusive
  decays, i.e. B?yKS
• We always need to find the 4-momenta of the
  particles at the origin of the decay
• This includes measuring 3-momenta of charged
  tracks and identifying them
• Also includes measuring energies direction of
  gs

6
Outline of Required Measurements

• How do we measure
• Charged particle positions momenta?
• Decay vertices?
• Gamma rays?
• Neutrinos?
• How do we identify charged particles (e, m, p, K,
  p)?

7
Trigger and Data Acquisition

• How do we acquire the data?
• Trigger All interactions are not interesting.
  Even at ee- colliders most of the collisions do
  not produce b-flavored hadrons. The largest rates
  are for
• ee- ?ee- (g)
• gg ? hadrons
• So a decision needs to be made quickly on
• writing out the data or losing it.
• DAQ The data acquisition system the hardware
  used for getting the data off the detector and on
  to tape, quickly.
• These considerations must be part of the overall
  detector design

8
Charged Particle Detection

• Charged particles are detected because they
  ionize electrons in atoms
• To calculate distribution of energy loss,
  consider elastic collisions of incident particle
  with atomic electrons
• Binding energy of electrons must be taken into
  account
• This is a difficult problem, first worked out by
  Landau Journal of Physics, 201 (1944)
• We need to know
• Energy distribution of ionized electrons (mean
  width)
• of electrons freed per unit length

9
The Landau Distribution

• Energy loss distribution is not Gaussian, has
  long tail toward higher energy
• Peak is called most probable energy loss
• Mean Energy loss is less well defined
• Many electrons involved usually

Most probable energy loss, Emp
Mean energy loss
10
Number of Produced e-

• N(dE/dx)/W where W depends on material (mean
  energy to produce an electron)
• Empirically determined
• 30 for 1 cm of gas (Ar, CO2)
• 25,000 for 300 mm of Silicon
• So detect charged particles by applying an
  electric field to collect electrons or some other
  means of seeing the ionization

11
Momentum Measurement

• Bend tracks in a magnetic field

• For q1, pt 0.3 r B,
• r in meters, B in Tesla, p in
  GeV
• For B 1T, and r 1m, pt 300 MeV

12
Bubble Chambers

• An old technology, no longer used
• Very illustrative
• Cold H2 liquid is both the target and the
  detector.
• Liquid is superheated boils due to ionization

13
Most Famous Bubble Chamber Event
14
Reasons bubble chambers are no longer used

• They cant cycle fast
• There is no electronic readout. The film must be
  scanned and then digitized by hand
• Particle identification is not good, nor is g
  detection (some experiments with lead plates)

15
Modern Tracking Detectors

• Gas detectors
• Detect ionization by applying an electric field
  to a thin (20 mm) wire
• Multiplication- when electrons get close to the
  wire they have enough energy to ionize the gas
  thus one ionization electron can turn into 10
  100 thousand detected electrons
• Geometry
• Many wires 50,000

charged track
e- drift towards the wire at constant drift
velocity
.
E1/r
16
Drift Chambers

• Position resolution
• In any measuring device of cell size s the
  accuracy in position is , determined
  by the s of a rectangular distribution
• Can do better by measuring the time from when the
  particle enters the system to when the first
  electrons hit the wire. This is called a "drift
  chamber." This works because the electron
  velocity in the gas is known
• Note 2-fold ambiguity

.
t
t
17
A Modern Drift Chambers

• The
• KLOE
• Drift
• Chamber,
• 50,000
• wires

18
Limits to Precision

• Magnetic Fields - inconvenient to get fields in
  excess of 1.8 T due to Fe saturation
• Finite time resolutions translate to real drift
  chambers having resolutions of 100-200 mm. Due to
  ionization statistics, wire position errors, e-
  drift velocity calibrations, etc..
• Multiple scattering- due to material

19
Stereo

• Problem How to measure two coordinates?
• Planar geometry crossed picket fences
• Cylindrical geometry small angle stereo, but
  this causes error to be much larger (10x) in one
  coordinate. Precision in
• r-f 150 mm, and in
• z 1.5 mm per layer

stero
axial
stero -
20
Multiple Scattering

• Its the material, stupid
• Due to Coulomb scattering from Nuclei. Well
  described by Moliere. Gaussian for small
  deflection angles, but with long tail due to
  Rutherford Scattering (This is a real pain!)
• For 98 of the scatters

where, Xo is the radiation length
21
How to Measure a Decay Vertex

• c b quarks are distinguished by their decay
  lengths. Lbgct, where t 1 ps.
• For bg p/m 1 t 1 ps, L 0.3 mm 300 mm
• Better if p is larger note tB 1.5 ps, tD
  1.1 ps
• First done by Framm at CERN 1982
• Made really good by E691 at Fermilab

22
Technology Use Silicon Detectors

• Silicon is made into a p-n junction diode with
  appropriate doping. It is operated at a bias
  voltage that forms a sensitive region depleted of
  mobile charge and sets up an electric field that
  sweeps charge liberated by radiation to the
  electrodes. 50 mm wide strips are placed on one
  side as a readout. Charged particles ionize the
  silicon and the charge is collected.
• Sensitive electronics are required
• Strips have relatively large Capacitance ? noise
• Many channels are required several 100,000
• Position resolution of 50 mm/?12 for binary
  readout, better for analog

23
Silicon detector picture
24
Problems with Silcon Strip Detectors

• Ambiguity problem
• Long strips are difficult because of large
  capacitance
• Lots of material on edges due to electronic
  readout

25
Pixel Detectors

• Make a square or rectangular array of silicon
• Put a small electronic circuit behind each one!
  (bump bonding)
• Send signals out to periphery where only hit
  pixels are readout "sparsification"
• Thicker than strip detectors, by a factor of 3
• Useful for high track density
• Useful for triggering on detached vertices

Each cell or pixel is small, ex 50 mm x 400 mm
26
BTeV Pixel Test Results

• Solid curve is a piece wise linear fit to a
  simulation based on a detailed Monte Carlo

Track angle (mr)
27
Photon Detection Electromagnetic Calorimetry

• Primary process is conversion of high energy g by
  pair production in Nuclear Coulomb field
• Process leads to a "shower"

atomic photoeffect
Rayleigh scattering
Pair production off Nucleus
Compton scattering
Pair production off atomic electrons
Photonuclear absorption
28
Sampling vs. Total Absorption Calorimeters

• A sample device uses a heavy material such as
  lead to convert the g's and then a sampling
  material such as plastic scintillator to "sample"
  the energy. The energy absorbed in the Pb is
  lost.
• Examples of sampling devices
• Pb-liquid Argon
• Pb-optical fiber (Shaslik)
• Typical Energy resolutions

29
Total Absorption Devices

• Idea here is to convert all the energy in an
  active medium
• Media can be transparent crystals CsI, PbWO4 or
  cryogenic liquids such as Krypton or Xenon (too
  expensive)

30
Calorimeter Readout

• Crystals
• Outside magnetic field photomultiplier tubes
  advantages fast, quantum eff 20, can be rad
  hard, but not cheap
• Inside magnetic field photodiodes, avalanche
  photodiodes (they have gain of 50), phototriodes
  (essential the first 3 stages of a pm tube)
  apd's are not cheap
• Cryogenic liquids collect charge on strips, so
  use analog electronics

31
Calorimeter picture
32
Neutrino Detection

• We cannot detect the small number of low energy
  neutrinos produced in b decays!
• n cross-section
• s(nN) 6x1039 cm-2 E(GeV)
• If anyone could figure out how to detect
  neutrinos it would make experiments much easier!

33
Charged Lepton Identification

• e use an electromagnetic calorimeter shower is
  almost identical to a photon
• m use the fact that muons don't have strong
  interactions. Use thick blocks of iron and see if
  the particles penetrate. Problem p and K decay
  into muons, so can get fakes

34
Charged Hadron Identification

• We are interested in separating p/K/p
• Technique depends on momentum range
• P lt 900 MeV Time-of-flight and dE/dx
• TOF equations
• dE/dx picture
• Some poor dE/dx info 1.5 3 GeV/c

35
Ring Imaging Cherenkov Counters

• Cherenkov radiation depends on particle velocity,
  sinqc1/nb, n is index of refraction
• Measure p using other devices so can derive m,
  the particle mass
• Many recent developments

36
Radiators

• Choice depends on velocity (or p range). Need
  nbgt1
• Require material to be transparent
• Desire low chromatic dispersion, i.e. n(E) to not
  be too bad
• In the few GeV/c range can choose liquids, ex.
  C5F12 or solids such as quartz or LiF

37
Photodetectors

• Must match radiator light output wavelength
  spectrum to that of photodetector
• Some possibilities
• TMAE gas 170-210 nm
• TEA gas 135-165 nm
• CsI (thin layer) 170-210 nm
• Phototube 250-550 nm

Use with wire chamber
38
Example of Comple System CLEO III RICH Detector

• Use CH4-TEA gas to detect single photons.
  Sensitive in VUV 135-165 nm
• Use LiF radiators
• Use N2 volume 15.7 cm thick to allow for
  Cherenkov cone to expand
• Use MWPC with pad readout to measure ? positions

39
One Cherenkov g Detector
40
Mating the Radiators to the Photon Detectors
41

• The End

42
Data Acquisition

• This arcane area is crucial in a modern
  experiment
• Functions include
• The trigger Which particle interactions do we
  read-out the detector? At Y(4S) ¼ of the ee-
  annihilations are b's, but there is a much larger
  rate of Bhabha and 2 photon collisions. In hadron
  colliders the b rate is much much lower, 1/500 at
  the Tevatron

43
DAQ continued

• Online monitoring Reads out monitoring signals,
  for HV gas, temperature, etcSamples the data
  during the run and histograms critical parameters
• Data Acquisition When a trigger occurs, the
  information from all the devices must be taken
  from the "front end" electronics and moved to
  "tape."
• Definition "Dead Time" The time that the
  experiment must be shut off to move the data onto
  tape. Modern readout systems try to eliminate
  this nasty feature.

44
dE/dx (mean energy loss)

• Note 1/b2 fall, and ln(bg) rise, called
  relativistic rise
• This is limited in materials by so called density
  effect, d term
• This information can be used to distinguish
  particles

45
Silicon strip Detectors

• Difference between
• spatial resolution
• impact parameter resolution
• decay length resolution

Impact parameter minimum distance of track from
a vertex
b
Spatial resolution inherent to the detector. For
50 mm strip width using binary (yes/no)
readout, r.m.s. resolution, s50/?12
Decay length distance Between primary
secondary vertices
L
46
Example of Complete System

• Delphi SLD Use Liquid radiators enclosed in
  quartz AND gas radiators. TMAE based wire
  chambers with long drift in E field