Transcript Strangeness Experimental Techniques (or how Theorists have it easy)

Strangeness Experimental Techniques
An expert is a man who has made
all the mistakes which can be made,
in a narrow field.-Niels Bohr
March 2003
Helen Caines
YaleUniversity
The Goal
To design the Ultimate Strangeness Experiment
What we need:
To be able to measure charged and neutral decays
Lots strangeness created.
Measurements at low and high pt.
Measurements at mid and high rapidity.
Only 5 charged particles are sufficiently stable to reach most
detector:
Pions, Kaons, Protons, Electrons and Muons
(+1) Photon: the only neutral particle which
can be efficiently detected.
SQM - March 2003
Helen Caines
V0s Reconstruction?
Strangely enough most
strange particles are neutral
or decay into something neutral
Strange Hadrons
K0S →p+p- (494 MeV/c2,
2.7cm, 68.6%),
L→pp- (1.12 GeV/c2,
7.9cm, 63.9%),
Ξ-→Lp- (1.32 GeV/c2,
4.9cm, 99.9%),
W-→LK- (1.67 GeV/c2,
2.5cm, 67.8%),
SQM - March 2003
Helen Caines
Finding V0s
proton
Primary
vertex
pion
SQM - March 2003
Helen Caines
Invariant mass distributions
•
Extracting the particle yields
– Consider each of the possible final states in turn
– Calculate the parent mass as a function of (y,pT)
M  m12  m22  2E1 E2  p1  p2 
– Count the number in the mass peak and correct for reconstruction losses
X
SQM - March 2003
Helen Caines
The Podolanski-Armenteros plot
Unique identification of two-body decaying particles by studying the
division, between the daughters, of the parent's mtm vector.
a - the fractional difference in momentum of the daughters
pt - the mtm component of the +ve daughter transverse to the line of
flight of the parent
p||  p||


pT  p  p a  
p||  p||
All possible values are constrained to lie along a curved locus specific
to the mass of the parent
a characterizes the decay asymmetry, <a> - daughter mass difference
SQM - March 2003
Helen Caines
In case you thought it was easy…
Before
After
SQM - March 2003
Helen Caines
Acceptance and Efficiency
SQM - March 2003
Helen Caines
Fine way to calibrate a detector!
• Large peaks at 2
o’clock and 8 o’clock
• TPC pad row
“floating”
SQM - March 2003
Helen Caines
Event mixing method
• The measurement of hadronic
resonances.
dN/dM [MeV-1]
Background subtracted K+K- invariant mass distribution
• These particles are short-lived
compared to the reaction time.
• Resonances are reconstructed
using a combinatorial technique.
Consider all track combinations
and calculate the invariant mass.
STAR Preliminary
f  K+K-
• The background is calculated using
positive tracks from one event
mixed with the negative tracks of
another event.
dN/dM [MeV-1]
K+K- invariant mass distribution (11% central events)
Same event distribution
Mixed event distribution
Invariant mass [GeV/c2]
SQM - March 2003
Helen Caines
Kink reconstruction
• In this method, one of the decay
daughters is not observed.
• Main background is from p-decay,
which has a smaller Q-value
• A cut on the decay angle
(momentum dependent) removes
the p contamination
• Remaining background from
multiple scattering and split tracks.
• Find ~ few kink decays per event.
Decay angle (degrees)
Approx. 10% of a real event showing a
reconstructed kaon decay K±  m ±n (64%)
or K±  p ±p0 (21%). Lifetime ct = 3.7 m
Kaon limit
Pion limit
SQM - March 2003
Helen (GeV/c)
Caines
Parent momentum
• So now we know how to reconstruct them.
•First question what kind of accelerator do we want?
SQM - March 2003
Helen Caines
Collider vs Fixed Target
Collider:
Fixed Target:
Higher energy
Lab frame == CM frame
Less focussed particles
Higher rates
Known z vertex
Boost gives longer ct
Go with the
SQMCollider
- March 2003
Helen Caines
Beam @ RHIC Complex
SQM - March 2003
Helen Caines
Beam at SPS Complex
The experiments Fixed Target Experiments
Forward emission at mid-rapidity
PS
North Area
NA
Booster
PS
LINAC3
ECR
SPS
West Area
WA
Experimental
Areas
Proton Synchroton
Fix target experiments
GeV
natPb ~ 4.25A
2
450
mg/cm
Heavy
Ion
LINAC
Electron Cyclotron Resonance
1mm
Al foil
PS
Booster
Super
Proton
Synchroton
Lead Plasma1mm
is 7bombarded
with
an e beam
carbon
foil
+82
+53
Pb
+82+27
PbPb
collisions
burst
Pb
95.4A
MeV
Pb10
160A
GeV
Pb+53 2.5A KeV per
Pb
SQM - March 2003
4.2A MeV
Helen Caines
Needle in the Hay-Stack!
How do you do tracking in this regime?
Solution: Build a high resolutiondetector
so you can zoom in close and “see”
individual tracks
Clearly identify individual tracks
Good tracking efficiency
SQM - March 2003
Pt (GeV/c)
Helen Caines
What’s the Main Tracking Device?
Advantages of a TPC (why there are 7 at RHIC)
• Large highly segmented tracking volume at low
cost
– Permits over sampling, a big plus
• Simplifies tracking code
• Improves position/momentum resolution
• Improves dE/dx resolution
• Design simplicity, an empty volume of gas
• Low mass – reduced multiple coulomb scattering
SQM - March 2003
Helen Caines
Disadvantages
• Slow readout – can’t be used in level 0
trigger
• Two track resolution limited with classic
design – (although improvements possible
with radial magnification)
SQM - March 2003
Helen Caines
How a TPC works
420 CM
•
Tracking volume is an empty volume of gas
surrounded by a field cage
• Drift gas: Ar-CH4 (90%-10%)
• Pad electronics: 140000 amplifier channels with
512 time samples
– - March
Provides
SQM
2003 70 mega pixel, 3D image
Helen Caines
SQM - March 2003
Helen Caines
Silicon Tracker?
Lots of material – Not so good – Lots of scattering
SQM - March 2003
Helen Caines
Charge Determination
Magnetic field
STAR, ALICE,
CMS, CERES,
NA49, NA57
Tracking detectors
NA50 & NA60,
PHOBOS,
BRAHMS,
ALICE
Trajectory
PHENIX,
In or Out of Magnetic Field
SQM - March 2003
Helen Caines
Charge transport correction (ExB)
In a non-uniform magnetic field
the drift velocity is not strictly
perpendicular to the pad-row
The Solution:
E  BB 2 2 

m  E B
v
2 2  E  B t  B  B  t 
1  t 

Find the drift velocity by solving
the Langevin equation


  

m v  q E  v  B  mAt 

At    v t , v  0
where:
qB

, m v E
m
A(t) is a stochastic damping term,
resulting from collisions in the gas
t is the mean time between
collisions
Horrible mess!!!!!!! Place in nice uniform Field
SQM - March 2003
Helen Caines
Now have Main detector
Want low momentum tracks , near primary vertex
Need fine pixelation
SQM - March 2003
Helen Caines
Fine Vertex Determination
ALICE ITS
SSD
SDD
SPD
Lout=97.6 cm
Rout=43.6 cm
6 Layers, three technologies (keep occupancy ~constant ~2%)
– Silicon Pixels (0.2 m2, 9.8 Mchannels)
– Silicon Drift (1.3 m2, 133 kchannels)
– Double-sided Strip Strip (4.9 m2, 2.6 Mchannels)
SQM - March 2003
Helen Caines
Position Sensitive Silicon Detector
Strip
Pad/Pixel
Drift
1eh/3.6eV, 300mm 25000 e
SQM - March 2003
Helen Caines
Charge cluster reconstruction
Cluster reconstruction
– Each pad-row crossing results in charge deposited in several
pad-time bins.
– These are joined to form clusters which have certain
characteristics
– The position is calculated as the weighted mean of the cluster
charge in the pad-time directions
– The coordinates are determined by:
• The mean pad position (x)
• The mean time bin x drift velocity (y)
• The pad row position (z)
– Total charge can be used for particle identification (see later)
SQM - March 2003
Helen Caines
Track reconstruction
•
Row: n-5 n-4 n-3 n-2 n-1
n
Step-by-step
– Find charge clusters in all TPCs
– Apply charge transport correction
– Track following in the main detector
• Start where the track density is lowest
• First find high momentum
seed
• Form initial 3 point tracks seeds
Track following method
• Use (local) slope to extrapolate the track
Row: n-5 n-4 n-3 n-2 n-1 n
– Tracks are propagated to inner detectors
• No momentum measurement outside magnetic field
• Assume all tracks are primary
• Momentum assignment based on trajectory
• Use trajectory to define a “road” in detector
– Search for non-primary vertex tracks
Track road method
• Do track following as a separate step
• Momentum determined from curvature of tracks
SQM - March 2003
Helen Caines
Calibration - Lasers
Using a system of
lasers and mirrors
illuminate the TPC
Produces a series of
>500 straight lines
criss-crossing the TPC
volume
Determines:
• Drift velocity
• Timing offsets
• Alignment
SQM - March 2003
Helen Caines
Momentum measurement
Measurement :
p0.3BTm GeV/c
Uncertainty:
 L s
2 

L
s
p B L s
 2 
2
 L
8s
L=3m, s=10cm
=11 m, B=0.5T
p=1.7 GeV/c
L
X+
P3
s
P2
p
B
L s
p B L
 2 8p s 2
p B
L Proportional
0.3BLto p
p  a  b p Constant
p
R
SQM - March 2003
P1
B
~ constant

dp e  
 vB
dt c
Helen Caines
Calibration – Cosmic Rays
Determine momentum
resolution
p/p < 2% for most tracks
SQM - March 2003
Helen Caines
Tracking Efficiency
•
Reconstruction losses can be divided into two types:
– Geometrical Acceptance
• Consequence of limited detector coverage
• Straightforward correction, calculated by Monte Carlo
– Reconstruction Efficiency
• Particles in acceptance but not reconstructed
• Possible reasons for loss:
– Hardware losses
– Detector resolution
– Merged/split tracks
– Reconstruction algorithm
– Efficiency correction
• Needs a detailed understanding of the detector response
– Embedding Method
• Tune Monte-Carlo simulation to reproduce the data
– Cluster characteristics
– Number of space-points on tracks
• Embed a few simulated tracks into real events
SQM - March 2003
Helen Caines
The Bethe-Bloch equation
• Energy loss
– Bethe showed that energy loss is strictly a function of b  v/c
and the properties of the medium
– Including relativistic effects, the Bethe-Bloch equation is
dE
1 2   2me c 2 2 2 

2
2Z

 4pN0 re mec  2 z ln
b    b2  
dx
A b
2

  I
where,
b  v c,   (1  b 2 ) 1 2
N 0  Avagadro Number


re  e 2 me , the classical electron radius
, Z, A  density, atomic and mass no.
I  ionisation potential
 parameterises saturation
SQM - March 2003
Helen Caines
Corrections
•
Experimental factors that affect the measured charge
– Temperature
• Controlled to better than ± 0.1o C
– Pressure
• TPCs are operated at “atmospheric” pressure
• Ionisation varies by 0.6% mbar-1
– monitored and normalised to 970 mbar
– Correction for O2 and H2O
• Both highly electronegative
• Results in linear charge loss with drift distance (few % in TPCs)
– Effective path length (dx) - depends on the track crossing angle
• Two angles: one in pad direction, the other in drift direction
– Equalise the electronic response
• Electronics and gas gain correction
– In practice an absolute gain calibration is difficult to obtain
• Inter-sector calibration (relative gain correction)
SQM - March 2003
Helen Caines
Comment on dE/dx measurements
•
Practical considerations
– All energy loss distributions have inherently large spread
• Primary ionisation
Number of primary electrons = DE /W
W = energy to produce e--ion pair
Follows Poisson statistics
• Secondary ionisation
Energy distribution of primary electrons ~ E-2
If E > W they can produce further ionisation
– Convolution produces Landau distribution
– TPC samples dE/dx from this distribution
– Use truncation to better estimate the mean
• Reduces sensitivity to fluctuations
• Typically drop ~20% highest dE/dx samples
• Truncation ratio must be optimised experimentally
– What happens in higher density media ?
 Fluctuations are reduced … but ...
 Height of rise decreases (probability for large DE increases)
 Momentum resolution worsens
(multiple
SQM
- Marchscattering)
2003
Helen Caines
Electronics and Gas Gain Calibration
•
Two methods
– Pulse the sense wires above the padrow
• Induces charge on all pads simultaneously
– Easy and quick to perform, but ...
– Measures electronics response at maximum load
– Doesn’t measure the gas gain
– Measure response to 83mKr added to the detector gas (ALEPH)
• Simultaneous calibration of electronics and gas gain variations
– 9 keV peak used to calibrate to MIP peak in data
– Provides linearity check up to several MIPs,
( depends on dynamic range of electronics)
MIP = Minimum Ionising Particle
SQM - March 2003
Helen Caines
PID Techniques(1) - dE/dx
12
Resolution:
dE/dx (keV/cm)
dE/dx
p
8
d
No calibration
With calibration 7.5%
Design
6.7%
K
p
4
9%
m
e
Even identified
anti-3He !
0
dE/dx PID range:
~ 0.7 GeV/c for K/p
~ 1.0 GeV/c for K/p
SQM - March 2003
Helen Caines
Still need high momentum PID
SQM - March 2003
Helen Caines
Time-of-Flight method
•
•
Requirements
– Time measurement between two scintillation counters (or similar)
– For p > 1 GeV/c, very good time resolution and long flight path
For example
– The time difference between two particles, m1 and m2, over a flight path, L, is
L
L
L
m12c 2
m22 c2 
Dt 

  1 2  1 2 
b1c b2 c c 
p
p 
which for p2 » m2c2 becomes
m12  m22 Lc

Dt ~
•
2 p2
NA49 Experiment
– The flight path is 13 m
– The time resolution st = 60 ps
– At 6 GeV/c: p-K separation = 2 st
K-p separation = 6 st
SQM - March 2003
Helen Caines
Now Have the Ideal Strangeness Detector!
A magnetic field for charge and momentum determination
A TPC for main tracking
An Innner silicon detector for high precision vertexing
and low momentum tracking
A TOF for high momentum PID
Tracking at high and mid-rapidity with large acceptance
Sound Familiar?
SQM - March 2003
Helen Caines
The STAR Detector
Magnet
Coils
TPC Endcap &
MWPC
Time
Projection
Chamber
Silicon
Vertex
Tracker *
FTPCs
ZCal
Endcap
Calorimeter
Barrel EM
Calorimeter
ZCal
Vertex
Position
Detectors
Central Trigger
Barrel
+ TOF patch
RICH
* yr.1 SVT ladder
Year 2000, year 2001, year-by-year until 2003,
SQM - March 2003
Helen Caines
installation
in 2003
Other Stuff
SQM - March 2003
Helen Caines
Energy Loss: Bethe-Bloch
SQM - March 2003
Helen Caines
Calorimeters
• Electromagnetic Calorimeters
• e- and  deposit their total energy in the Calorimeter
• Hadronic calorimeter (may be in the future at mid-rapidity)
• Zero Degree Calorimeters are largely used
• High Multiplicity :
• Small RM ~ 2-5 cm
• Distance 4-5 m from IP
• Spectrometer
• Sampling Calorimeters:
• cheap (acceptance)
• Lead+Scintillator
• Homogenous Calorimeters :
• Resolution,
• LeadGlass, PbWO4
PbWO4
X0 0.89
RM 2 cm
lI 19.5 cm
n 2.16
Res 3% at >3GeV
PHOS in ALICE & ECAL in CMS
SQM - March 2003
Helen Caines
Triggering/Centrality
• “Minimum Bias”
~30K Events
down the beam line |Z | < 200 cm
vtx
Participants – Definitely created moving away from beamline
ZDC East and West thresholds set to
Spectators – Definitely going
lower edge of single neutron peak.
REQUIRE:
Coincidence ZDC East and West
• “Central”
CTB threshold set to upper 15%
Several meters
REQUIRE:
Min. Bias + CTB over threshold
Spectators
Impact
Parameter
Participants
Zero-Degree
Calorimeter
Spectators
SQM - March 2003
Helen Caines
RHIC ZDC
SQM - March 2003
Helen Caines
V0 Efficiency
Decay Length
DCA to PV
Neg. Daughter
DCA
Daughter DCA
Pos. Daughter
DCA
Mult.
SQM - March 2003
Helen Caines
Kinematics
Eventm, p,,
N
i0
s
d
E p
d
3
Invariant
cross-section
3
Lorentz Transformations
 E*    b 0  E 
 *
 
 p//  b  0  p// 
 p*T   0 0 1  pT 
  
 
Eventm, pT, y,
N
i0
ds
2p pT dy dp
2
T
SQM - March 2003
Helen Caines
Why Rapidity?
Kinematical reason:
•The shape of the rapidity distribution, dn/dy, is invariant
 E p z 

y0.5ln
 E p z 


y* = y + y0
Dynamical reason:
• The invariant cross-section can be factorized
s
n
ds
d
d


2p pT dy dp 2p pT dp dy
2
T
T
SQM - March 2003
Helen Caines
Pseudo-Rapidity

 
 
  ln tan 
2
 
y, pm,  1/
yp /2
SPS =9, >>6o // RHIC =100, >>1.6o // LHC =2750, >>0.02o
Maximun Rapidity
 s
ymax ln m 
 
SPS
ymax 2.8
RHIC
ymax 5.3
LHC
ymax 8.6
SQM - March 2003
Helen Caines