measuring cdf’s run ii luminosity
DESCRIPTION
Measuring CDF’s Run II Luminosity. Introduction Tevatron Luminosity measurement techniques The CLC Detector Gas, cones and light collectors PMTs Mechanical design Simulation Event generator Detector Test beam Performance CDF luminosity measurement Summary - PowerPoint PPT PresentationTRANSCRIPT
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Measuring CDF’s Run II Measuring CDF’s Run II LuminosityLuminosity
Introduction Tevatron Luminosity measurement techniques
The CLC Detector Gas, cones and light collectors PMTs Mechanical design
Simulation Event generator Detector
Test beamPerformance
CDF luminosity measurementSummary
Roberto Rossin for the CLC group
University of Florida
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Intro
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Collider Beam Collider Beam LuminosityLuminosity
Collider Beam Collider Beam LuminosityLuminosity
Lttppn )(events) top( EfficiencyAcceptanceBR
diameter) beam 100(~ cm 103~
spacing)bunch sec (132 MHz 2~
(36) ringin bunches ofnumber
) 104~(nch protons/bu-anti
) 103~(nch protons/buN
242
0
10
11p
nf
B
N p
Total Lum:
=Np Np Bf 0
42L
Instantaneous Luminosity:
~ 2x1032 cm-2s-1
(Run IIa goal)
dt L L goal) IIa (Run -1
fb 2 ~
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
36x36 Bunch Structure36x36 Bunch Structure
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Run II Tevatron LuminosityRun II Tevatron Luminosity
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Run II Integrated Run II Integrated LuminosityLuminosity
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Run II: now and in 2009Run II: now and in 2009
980 GeV, 36 x 36
Parameter Now 2009 _
Initial Luminosity 120-160 270 E30 cm-2 s-1
Integrated Luminosity 16 28 pb-1/week
Tot. Int. Luminosity 1.6 4.8-5.4 fb-1
Protons/bunch 235 300 E9
Antiprotons/bunch 35-50 75 E9
Proton emit. (95%,norm) 16 16 mm-mr
Pbar emit. (95%,norm) 11 8 mm-mr
Beta @ IP 0.28 0.28 meter
Hourglass factor 0.58 0.58
Peak Pbar Production Rate
16.6 >20 E10/hour
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Luminosity Detector in Run IILuminosity Detector in Run II
Operate at: L ~2*1032cm-2sec-1 7ppbar/b.c.
Run I Run II
L
Measure instantaneous and integrated luminosity for CDF
and Tevatron in real-time (~ 1 Hz),
standalone, and offline for physics
delivered and live luminositybunch by bunch luminositykeep good precision at high
luminosity: < few % Measure z profile of collisions Provide a minimum bias
trigger for CDF. Diffractive trigger NEW: veto for b-physics
triggersLinear, robust, and with many handles for very high Luminosity
CLC
Finding room in CDF IIFinding room in CDF II
Plug Upgrade
Central Muon Upgrade
Central
Muon
Extension
Silicon Vertex
Detector
Forward Muon Upgrade
Low Beta Quad Low Beta Quad
Low Beta Quad Low Beta Quad
Run I
Run II
Intermediate
Muon System
plug
beam-beamcounters
10o
3o
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Pointing Gas Cherenkov Pointing Gas Cherenkov CountersCounters
Pointing Gas Cherenkov Pointing Gas Cherenkov CountersCounters
Sensitive to the right particles! Much light for particles from interactions Little light from secondaries and soft
particles
• Cherenkov thresholds
• Shorter paths Not too sensitive to particles from back
(halo) Excellent amplitude resolution
Count # hits and # particles No saturation nice linearity
Excellent time resolution Distinguish # of interactions by time
Robustness Radiation hard / low mass
Disadvantages: Needs gas system
New idea, more interesting....
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLCTechniques for Luminosity Techniques for Luminosity
measurementsmeasurementsTechniques for Luminosity Techniques for Luminosity
measurementsmeasurements
Use a relatively well known, copious, process: Inclusive inelastic p-pbar cross-section
large acceptance at small angles
Lf in
• µ = avg. # of interactions/b.c.
• f = frequency of bunch crossings
• = tot inelastic cross-section
• L = inst. luminosityin
Use a good estimator Measure the fraction of crossings with p-
pbar interactions (Run 1 “rate” technique)• Use: prob. of no int.
Direct counting # of p-pbar interactions Counting particles Counting time clusters
Cross-calibrate with rarer, clean, better understood process:
Need full understanding of tracking, particle-id, missing-Et,, trigger, NLO, bcknds etc. Useful for integrated lum abs. normalization
,lepW
det ectorBC inf L
Use dedicated detector:
eP0
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
60.40 (1.40)
81.90 (2.30)
CDF extrap.@ 2.0 TeV
55.92(1.19) 2%
60.33 (1.40) 2%
71.71 (2.02)
80.03 (2.24)
E811Exp.
CDF meas.
@ 1.8 TeV
Process (mb)
8%in
tot
The total pp cross sectionThe total pp cross section
Elastic Scattering Single Diffraction
M
Double Diffraction
M1 M2
Hard Core
Proton AntiProton
PT(hard)
Outgoing Parton
Outgoing Parton
Underlying Event Underlying Event
Initial-State Radiation
Final-State Radiation
“Hard” Hard Core (hard scattering)
Proton AntiProton
“Soft” Hard Core (no hard scattering)
+ ++
+
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Luminosity methodsLuminosity methods
Measure energy deposited on CLC Particle counting Count # of channels above threshold Hit counting (Implemented)Count empty events Zero counting (main method so far) Count # of clusters in time Time clusters counting
Need to calibrate detector.. Need to calibrate detector.. estimateestimate
which accounts for limited which accounts for limited coverage and efficiencycoverage and efficiency
From simulationsNeed all relevant material in CDFNeed “correct” generator…
From real dataW’s, Z's, J/'s yields
Particle counting with thresholds
Hit counting
Zero counting
Particle counting
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
“particles” hitsZero bias data SimulationPerfect linearity
expected L in Run II
Particle counting – Hit counting Particle counting – Hit counting methodsmethods
1P
P
in
BC
N
NfL
= avg. # particle for a single p-pbar interaction.
= measured avg. # particle/bunch crossing
1PN
PN
1H
H
in
BC
N
NfL
= avg. # hits for a single p-pbar interaction.
= measured avg. # hits/bunch crossing
1HN
HN
Measured at low luminosity from 0-bias data
Hit Counting
Particle Counting
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLCParticle counting – Hit counting Particle counting – Hit counting
methodsmethods
Hit Counting
Particle Counting
Advantages:No saturation @ high LLinear with physical backgrounds
Disadvantagesgain/gain= L/L
Advantages:Less sensitive to gain/gain
DisadvantagesNot linear with physical backgroundsSaturation @ high L
Define a collision { >0 hits in E} .and. { > 0 hits in W} with amplitude > Ao
• Online has fixed thresholds• Offline we can normalize to single particle peak (spp)
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Threshold
Luminosity with zero counting methodLuminosity with zero counting method
Zero Counting (empty crossing)
Amplitude (p.e.)
single interaction
five interactions avg.
1 part.
1 part.
2 part.
100 p.e.secondaries
primaries
Measure the average of the poisson distribution by measuring the 0 bin0 bin.
An empty crossing is a BC without E–W coincidence in CLC.
eP )(0
Advantages:Less sensitive to gain/gain Formula is correct at every L
DisadvantagesNot linear with physical backgroundsStatistically limited at very high L
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Luminosity counting time Luminosity counting time clustersclusters
Time clusters
interaction time
~100ps
=1.4ns0
1
2
3
4
1 2 3 4 5
counter arrival times
Measure the number of p-pbar interactions using
precise timing
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Detector
design
CLC
CDF II cross-sectionCDF II cross-section
0
.5
1.0
1.5
2.0
0 .5 1.0 1.5 2.0 2.5 3.0
300
SOLENOID
= 1.0
= 2.0n
= 3.0n
n
m
m
END WALL HADRON CAL.
CENTRAL OUTER TRACKER
COT Endplate
ISL
SVX II
Electronics
(m)Z
Plug EMPlug Had.
CLC
CENTRAL CALORIMETER
10o
3o
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CLC Mechanical CLC Mechanical DesignDesign
48 counters/side3 layers with 16 counters eachrapidity coverage 3.7< <4.7Isob pressure: up to2atm
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CLC location in CDFCLC location in CDF
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CLC DesignCLC Design
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Cherenkov radiation cos = 1/( n) Npe = No L <sin
No = 370 cm-1 eV-1 col(E)det(E) dE
col - light collection efficiency
det - PMT quantum efficiency
Pions > 2 GeV & Electrons > ~9 MeV
Isobutane @ 1atm n=1.00143 =3.1o
sin col>~0.80x0.80 = 0.64 det(E) dE ~ 0.84 (quartz
window) No ~ 200
Np.e.~110 (L=200cm)
Cherenkov Counters – Cherenkov Counters – prototypeprototype
Cherenkov Counters – Cherenkov Counters – prototypeprototype
particleL
PMT
mylar cone light collector
window
cm 5~
gas
1 / nLight emitted if UV dominated
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Gas & conesGas & cones
Gas: Isobutane at low (< 2atm) pressure Higher pressure:
higher refraction index higher amount of light higher Cherenkov angle more reflections
Cones Aluminized mylar:
2 x 0.1 millimiters thick 60 nm aluminum layer MgF2 coated reflectivity for Cherenkov light > 90%
Inner Layer Middle Layer Outer Layer
d 0.831 in 0.854 in 1.211 in
D 1.175 in 1.653 in 2.318 in
L 43.498 in 71.813 in 70.610 in
0.45 degree 0.64 degree 0.90 degree
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Aluminized Mylar cones - designAluminized Mylar cones - design
2cos)(
/2
2/
minmax
minmin
c
c
RRL
dR
Light collector (Al)
cone cap2 layers; 0.1 mm each; 60 nm Al coating
(rad hard)
50 nm Al + 50 nm MgF2
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Light collectorsLight collectors
Light collectors Used to collect light onto the face of the
PMT Aluminum with conical shape
optimized for maximum light collection efficiency
inner: single cone middle & outer: bi-cone (10° and 14°). For
large collectors bi-cone has higher efficiency and uniformity. Give a less divergent light beam
Polished to optical quality; coated with 50 nm Al and 50 nm MgF2.
Gap between collector and PMT window allow to hide PMT deep inside magnetic shield. L/D > 1 in order to protect PMT against the
magnetic field (L=distance between shield edge and PMT, D=shield diameter)
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Light collection efficiency Light collection efficiency simulation simulation
Cone+coll efficiency VS collector-PMT gap
Collector efficiency VS photon angle
gap = 0, 5, 10, 15 mm
Collector eff:bi-cone single cone
Total eff:bi-cone single cone
Cone eff VS photon position
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Geometrical constraints ~ 25 mm diameter (all layers…)
Larger doesn’t fit Smaller requires large angle
collector (losses)
Performance Quartz window
UV transparent Rad hard Thin
Less light Better timing resolution
Timing resolution Sub 100 ps resolution
PMT choicePMT choice
R5800Q Hamamatsu10-stage / 106 gain
0.8 mm concave-convexquartz window25 mm x 60 mm1.7 ns rise-time
~12 ns pulse width
PMT
R5800Q parameters
mm 25
cathode mm
21
N. of stages
10
Q.E.% 27
Voltage, kV 1.8
nominal gain, 106
2.0
dark current, nA
2
dark noise, kHz
6
rise time, ns
1.7
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
CLC systemsCLC systems
CDF
CLC East CLC West
CCCooolllllliiisssiiiooonnn HHHaaallllll CCCooouuunnntttiiinnnggg///CCCooonnntttrrrooolll RRRoooooommmsss
CLCFront EndElectronics
Crate
CLC OnlineLumi DAQ
CLC LEDCalibration
System
HV CAENSY527/A932N PC
CDF TriggerSupervisor
CDF DAQ
Ethernet toTevatron
70 m coax cable
70 m coax cable
70 m coax cable
Min. BiasCoincidence
Unit CDF Level-1trigger
Gas System
CLCLuminosity Readout & Luminosity Readout &
Information FlowInformation Flow
CLLD
CLLD
OFFLINE ONLINE
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Calibration Calibration systemsystem
white Tyvec
quartz fibers
• Green-blue bright LED from Ledtronics• Quartz fiber + Tyvec reflector• Pulse generator @ 8 V x ~40 nsec duration • 100-500 mV signal for gain calibrations• About 600 psec r.m.s. signal •Use peak for initial timing calibration
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Full detector
simulation
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Simulation: event generator Simulation: event generator
ppbar processes: elastic scattering (pp pp) diffractive scattering (pp pX,pX,X1,X2) "hard core" scattering (pp X)
for luminosity only inelastic
Event generators: MBR Pythia
both have been tuned to the CDF data typical multiplicity in one ppbar primary
interaction is 5 particles per unit of . luminous region has gaussian shape in z
and t.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Simulation: CDF detector and CLCSimulation: CDF detector and CLC
CDF detector simulation based on Geant 4 simulation accounts for all secondary
particles with energy > 5MeV total number of particles detected is
considerable larger than the primary ones ~50% from IP 25% from beam pipe 25% from plug calorimeter
Gas counter simulation light yield accounts for:
the geometrical path the reflectivity (90%) threshold of the Cherenkov radiation.
-electron are below thr.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Simulation: Momentum Simulation: Momentum ThresholdThreshold
Simulation: Momentum Simulation: Momentum ThresholdThreshold
Simulation accounts for the momentum threshold
for isobutane at 1atm threshold is for = 19 2.6 GeV pions 9.5 MeV electrons
distribution almost flat at ~ 20: 10% pressure change = 4% th change = 0.8% N particles
factor
momentum (GeV)
primaries
secondaries
all
th=20
pions ~ 2 GeV
Cherenkov p threshold
electr. ~ 8 MeV
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
All particles
Particles from PV
All secondaries
Secondaries from plug
Simulation: amplitude Simulation: amplitude distributiondistribution
Simulation: amplitude Simulation: amplitude distributiondistribution
Simulated amplitude for particles from minimum bias pp interactions.
primary particles 100 p.e. peak low amplitude tail due to parallax
secondaries have low amplitude 50 p.e. threshold cuts:
84% secondaries 23% of primaries
Simulated amplitude for 5 pp interactions. primary particles
single and double particle peaks particle counting can be
performed secondaries still have low amplitude
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
PMT window & electronsPMT window & electrons
Fraction of particles that hit window light yield set to 26p.e./mm window thickness = 5mm Overall effects:
peak width increased by 20% peak shift: 111 p.e. 115 p.e.
Electrons contribution mayor contribution
to secondaries electrons produced
on beampipe give large signal
all particles
hit window
electrons
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Detector
performance
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
First collisionsFirst collisions
First collisions at 1.96 TeV at CDF seen by Cherenkov Luminosity Counters in October 2nd 2000
1 proton bunch vs 4 anti-proton bunches 19 ns separation between
anti-proton bunches
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Beam structure: timingBeam structure: timingT
ime
(ns)
Time (ns) Time (ns)
Collisions main bunch
Collisions satellite bunch+ losses
396 ns
Losses
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
P-pbar interactions Z positionP-pbar interactions Z position
From CLC timing information determine Z position of collisions
• Was VERY useful during Tevatron commissioning:
•CLC was the only detector which could measure position of interactions during that time
•Helped a lot to AD to tune the machine
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
CLC amplitude distributionCLC amplitude distribution Single Particle Peak (spp)
Higher CLC occupancy due to more material close to beam SPP clearly seen after isolation cat useful for calibration Great agreement with simulation
amplitude, spp units
amplitude, spp units
Simulation (dots)East module (blue)West module (red)
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
T0 CalibrationT0 Calibration
•Calibration using p and pbar beams
Time offset
w.r.t. a reference channel
(before applying
T0 correction)
WESTWEST
EAST EAST
Time offset
w.r.t. a reference channel
(after applying
T0 correction)
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
LuminosityMeasurements
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Luminosity with zero counting methodLuminosity with zero counting method
1)( )1(0
0 WE eeeP
0 = probability of NO hits on both modules
E/W = probability of hits only on East/West module
Provided by Monte CarloProvided by Monte Carlo
Zero counting (empty crossing)Zero counting (empty crossing)
Measure the average of the poisson distribution by measuring the 0 bin0 bin.
An empty crossing is a BC without E–W coincidence in CLC.
In an ideal world (…acc=4 and =100%) eP )(0
…But life is tough
0 7%
E/W 12%
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Empty crossing methodEmpty crossing method
Empty crossings
Hit Counting
D0
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Real Time LuminosityReal Time Luminosity
Delivered Inst. Lum
CDF live Inst. Lum
Proton losses
Abort gap losses
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Bunch by Bunch LBunch by Bunch L
Instantaneous Total
Instantaneous Total
Tevatrondelivered
CDFlive
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC WWllXsecs & Z Xsecs & Z ll Xsec ll Xsec
Measured with 194 pb-1
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC WWee yield versus luminosity yield versus luminosity
Data collected from Jun 23 to Jul 3 2004
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC J/J/ yield versus luminosity yield versus luminosity
Data collected from Jun 23 to Jul 3 2004
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CDF vs AD LuminosityCDF vs AD Luminosity
0
20
40
60
80
100
120
0 20 40 60 80 100 120
CDF vs AD Luminosity
CDF Luminosity
AD Luminosity
LCDF
/LAD
= (1.03 ± .03) -(.0004 ± .0004)LAD
• AD calculation from Super-table (p, pbar, ’s)• Slope compares AD estimate of * with that observed at CDF
Data collected from May 17 to Jul 16 2004
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CDF Luminosity vs COT currentsCDF Luminosity vs COT currents
Data collected from Jul 18 to Jul 31 2004
Removed the first hour on each store to have stable COT currents.
3%
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
ConclusionsConclusions
Established all Lum measurements and accounting
Uncertainty below 5% level Used W’s for cross-check
Looks good
Implemented offline stability tests Implementing high lum algorithms
Particle counting Time clusters
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
CLC PublicationsCLC Publications
“Luminosity Monitor Based on Cherenkov Counters for P Anti-P Colliders” Nucl. Instr. & Meth. A441 (366-373), 2000.
“The CDF Run II Luminosity Monitor”
Nucl. Instr. & Meth. A461 (540-544), 2001.
“The performance of the CDF Luminosity Monitor”
Submitted to Nucl. Instr. & Meth. May, 2002.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Backup Plots:
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Test beam results
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Cerenkov lightCerenkov lightCerenkov lightCerenkov light
The number of photoelectrons in a counter:
N p.e. L 2z 2
remec2 coll(E)det(E) sin2 c(E)dE
2z 2
remec2 370 cm -1 eV-1
n is usually const. over the useful range of photocathode sensitivity, therefore:
N p.e. 370L sin2c coll(E) det(E) dE
Using det(E) dE 0.27sin2 c 2(n 1)
N p.e. 200 n 1 L cm coll
n 1(no 1)P
Po
Pressure dependence:
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
150 GeV
TB setupTB setup
Beam: 150 GeV pions Timing & triggering: 2 scintillating counters
upstream and downstream Drift chambers for particle position detection Alignment precision ~1mm
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
TB setupTB setup
Cherenkov luminosity monitor prototipe Cone:
2 m long 50 m thick 2.0 – 4.3 cm openings 5.0 cm long light collector
Gas: 6 different choices pump: 0.1 – 2 atm
PMTs: 5 choices
19-53 cm LED:
light yield calibration
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
TB setupTB setup
Electronics and data acquisition Purpose: measure amplitudes and
time differences between the counters
Feed signal to dividers. 10% for charge measurement.
Amplified 10x, delayed 250ns. sent to ADC
linearity better than 5% gated (100ns) by trigger
from scintillators 90% for time measurement.
feed linear discriminators Up scintillator starts TDC,
(50ns delayed for monitoring) Down scintillator and
Cherenkov stop TDC NIM coincidence unit to form
a trigger
Sc. Up
Sc. Down
Cher.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC TB: amplitude TB: amplitude measurementmeasurement
To count photoelectron in the Cherenkov signal we need to know Np.e. /ADC.
Blue LED inside the gas volume near the entrance of the cone.
Amplitude calibration - 1 pulser at very low amplitude
ensure single photoelectron signal from PMT split:
low threshold discr. to form a trigger with the pulser
to ADC (gated with 100 ns from trigger)
Amplitude calibration – 2 Width of the amplitude spectrum
which has a gaussian form determined by statistics of photoelectron.
2p.e 1p.e.N 1
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC TB: amplitude TB: amplitude measurementmeasurement
Single p.e. Multi p.e.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC TB: amplitude TB: amplitude measurementmeasurement
Amplitude calibration results
parameter HM R2076
HM R5800Q
HM R7057
Ph XP2978
Ph XP2020Q
, cm 19 25 29 29 53
Voltage, kV 1.8 1.6 1.6 1.4 1.7
single p.e , p.e. ~0.5 0.42 0.41 0.45 0.50
single p.e. calibration, p.e./count
~4.5 3.7 4.5 1.4 1.8
multi p.e. calibration, p.e./count
4.0 3.2 4.0 1.2 1.8
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
TB: detector alignmentTB: detector alignment
Alignment provided by DC located upstream: PMT window has variable thickness, light yield in there changes with x,y. Light yield in the cone also depends on x,y. Triggering on Cherenkov signal ( PMT = 53mm)
Before alignment After alignment
PMT window light
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
TB: detector alignmentTB: detector alignment
Before alignment After alignment
PMT window light
Cone light
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC TB: light yield from gas TB: light yield from gas and PMT windowand PMT window
Want to separate contribution of gas from the one from quartz window. Fill tube with N2 at 0.11atm and
measure light yield Fill tube with the test gas at
1.47atm, repeat measurement Keep PMT HV up for stability
Results for XP2020Q ( = 53mm) window yield increases with radius
(thickness not constant), in the center Np.e.=52 Np.e./mm=26
after correcting for pressure, isobutane has Np.e.=1225 @ 1atm, this corresponds to N0=214 p.e./cm
C4H10 – 1.47atm
N2 – 0.11atm
PMT gas only yield
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC TB:light yield from gas. TB:light yield from gas.
Pressure dependence isobutane
Different radiators
gas iso-C4H10 C2F6 C3F8 C4F8 SF6 N2
measured yield 100% 61% 79% 86% 58% 23%
expected yield 100% 55% 74% 87% 64% 21%
Oxigen contamination: nitrogen
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC TB:light from small PMTs TB:light from small PMTs
Collector efficiency tested on PMT with =53mm = 86%
Light yield in C4H10 for R5800Q
Light yield in window VS radius: measurements in 0.11atm N2.
C4H10 – 1.47atm
N2 – 0.11atm
PMT gas only yield
=53mm
=19mm
=25mm
=29mm
=29mm
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Detector assembly
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Outer layer cone + collectorOuter layer cone + collector
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
CLC inner structure assembly (at UF)CLC inner structure assembly (at UF)
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Quartz fiber bundles: LED Quartz fiber bundles: LED calibration systemcalibration system
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC The CLC modulesThe CLC modules
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Completed PMT assemblyCompleted PMT assembly
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Completed modulesCompleted modules
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
LuminositySystematics
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Systematics for zero countingSystematics for zero counting
L
L
L
LL
''
)',',','(),,,( 0000 WEWE PP
Given a “baseline” set of (0, E, W)
and a new set of ('0, 'E, 'W ) both evaluated from MC
Solving
for ' we get the sys on L
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Systematics: uncertainty sourcesSystematics: uncertainty sources
Inelastic p-pbar XsecInelastic p-pbar Xsec
CLC acceptance:CLC acceptance:
Geometry & generator
Beam position
CLC simulation
SPP calibrations
Acceptance stabilityAcceptance stability
Online Online offline transfer offline transfer
Beam lossesBeam losses
Statistical uncertaintyStatistical uncertainty
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Systematics:Systematics: Geometry and generator Geometry and generator
Generator is Pythia (Tune A).Generator is Pythia (Tune A).Hard core + single diffractive + double diffractiveHard core + single diffractive + double diffractive
CDF simulator is Geant4CDF simulator is Geant4Modified effective thickness of beam pipeModified effective thickness of beam pipe
CLC simulation, changed the pressure of CLC simulation, changed the pressure of IsobutaneIsobutane
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Geometry and generatorGeometry and generator
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Systematics on luminosity due to detector simulation and generator have been evaluated to be in total ~ 3% ~ 3% (=3(=32%)2%).
Geometry and generatorGeometry and generator
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Systematics on luminosity due to CLC simulation have been evaluated to be ~1%~1%.
CLC simulationCLC simulation
particleL
PMT
mylar cone light collector
window
cm 5~
gas
Photons propagated one by one. Photons propagated one by one.
Response depends on:Response depends on:Gas pressureGas pressureMylar reflectivityMylar reflectivityPMT efficiencyPMT efficiency
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Systematics from Beam PositionSystematics from Beam Position
We split the RunII into 7 intervals with stable beam configuration in term of:
X0 ,Y0 and
X/Z, Y/Z (slopes).
We are not sensitive to Z0.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLCSystematics from Beam PositionSystematics from Beam Position
Systematics on luminosity due to beam position is < 1%< 1%.
In every interval the beam configuration affects luminosity by less than 1%.
Taking the average over the whole RunII the effect is negligible.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLCCLC in situ calibrationCLC in situ calibration
Store 3530 (May 25)
Init. Lum=68.4*1030
Pedestal fit
Single Particle Peak default fit
Two particle peak
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Systematics from SPPSystematics from SPP
ISO cut Lum dep. Fit meth.
PED syst. TOT
Systematics ON SPP 2.1% 1.2% 1.3% 1.5% 3%
See Sasha P. talk
single interaction
five interactions avg.
1 part.
1 part.
2 part.
100 p.e.secondaries
primaries
Threshold
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Systematics from SPPSystematics from SPPWorst case scenarioWorst case scenario
Systematics on luminosity due to SPP uncertainty is < < 1%1%.
At very low gain regime, the impact on luminosity is more significant.
Store 2606
See Sasha P. talk
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Contribution due to lossesContribution due to losses
We consider one store with large variation of proton losses, due to a rescraping during the store itself.
We fit the online luminosity measurement in the two regions and calculate the JUMP before and after rescraping.
Delta luminosity is an overestimation of the beam losses contribution (luminosity NEVER increases during rescraping).
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC May 16 2004, store 3500May 16 2004, store 3500
Between 04:45 and 05:30 proton losses had been permanently above 70kHz.
Beam rescraping at 05:33
After rescraping, losses where below 20kHz.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Titolo:psfile.ps: Lum VS LossesAutore:ROOT Version 3.10/02Anteprima:L'immagine EPS non è stata salvatacon l'anteprima inclusa in essa.Commento:L'immagine EPS potrà essere stampata con una stampantePostScript e non conaltri tipi di stampante.
Contribution from beam lossesContribution from beam losses
Contributions to luminosity due to beam losses is < 1%< 1%.
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Titolo:
Autore:
Anteprima:L'immagine EPS non è stata salvatacon l'anteprima inclusa in essa.Commento:L'immagine EPS potrà essere stampata con una stampantePostScript e non conaltri tipi di stampante.
Statistical uncertaintyStatistical uncertainty
For NTOT=1 ii < 4 =
Statistical uncertainty per bunch crossing is
044.08243
4:
i
i
Statistical uncertainty per single measurement (36 bunches) is
Statistical uncertainty on luminosity is negligible.negligible.
007.036
044.0
ML
L
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC Uncertainties summaryUncertainties summary
L/L is 4.2L/L is 4.24%=4%=5.8%5.8%
Lum Range
CLC acceptance:
Geometry <3% 2x1032
Generator <2% 2x1032
Beam <1% 2x1032
CLC simulation <1%
Spp calibration <1% 2x1032
Acceptance stability 1% --Online offline transfer -- anyBackgrounds <1% --Non linearity@high L --
Statistical uncertainty -- 2x1032
Inelastic p-pbar Xsec 4% --
TOTAL <4.2%
Systematic Error
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Cerenkov lightCerenkov lightCerenkov lightCerenkov light
The number of photoelectrons in a counter:
N p.e. L 2z 2
remec2 coll(E)det(E) sin2 c(E)dE
2z 2
remec2 370 cm -1 eV-1
n is usually const. over the useful range of photocathode sensitivity, therefore:
N p.e. 370L sin2c coll(E) det(E) dE
Using det(E) dE 0.27sin2 c 2(n 1)
N p.e. 200 n 1 L cm coll
n 1(no 1)P
Po
Pressure dependence:
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
CLC timingCLC timingCLC timingCLC timing
Precise Timing measurements: Use 1ns TDC (standard CDF) Use time stretcher circuitry
~ x 10 stretching factor
sub 100 ps resolution
CLCDiscriminator
in Transition Board Stretchers TDC
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Luminosity FormulaLuminosity Formula
The major luminosity limitations are•The number of antiprotons ( )
•The proton beam brightness (Np/p)
•F<1
BNp
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
CLC parallaxCLC parallax
int. regionbeam
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Magnetic ShieldingMagnetic Shielding
CLC
Magnetic ShieldingMagnetic Shielding
“pre-shield” optimization
Bz
(G)
Z (cm)
IndividualMu-metal shields
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Cerenkov - 48 chann.Cerenkov - 48 chann.Cerenkov - 48 chann.Cerenkov - 48 chann.
Some sources of systematics: Particle multiplicity/interaction Fraction of secondaries Source of secondaries PMT window thickness Gas index of refraction
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC
Reference ProcessesReference Processes
Process of inelastic PPbar scattering Large x-section: (CDF)
Total x-section is measured also by E710 and E811 (2.8 discrepancy with CDF)
Process: W l Complementary L measurement with different systematic
error Cross-section ~2.5nb (well known
theoretically:20%NLO,3%NNLO) Expected rate (@L=2 1032) 0.5Hz (good for integrated L) Trigger&selection efficiency ~25% (rate of ~0.1Hz after cuts)
LfR clcinelBCpp L – luminosity– # of interactions /BCinel – inelastic x-sectionclc – CLC acceptance fbc – Bunch Crossing rate
LR lWlWlW
mbinel 4.14.60
R = rate
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CLC effective inelastic x-CLC effective inelastic x-sectionsection
Method
CLCX-sec (mb)
2 layers
CLCX-sec (mb)
3 layers
Old simulationMBR + 3 Xo
Straight acceptance
36.9 40.5
Old simulationMBR+3 Xo
CLC/(CLC+plug)35.2 38.6
Data (CLC+plug accept from old
sim.)CLC/(CLC+plug)
35.8 38.3
New SimulationStraight
acceptance38.2 42.4
Notes: Span in x-sec is 10% So maybe error is ~ +-5 % ? Needs more time to settle and
to decide what’s “more correct” MBR inelastic is 44.4 mb ~
measured by CDF Pythia gives ~9% smaller CLC acceptance for hard-core
(@500 ADC) is ~90%. Is ~95% for lower threshold.
Need to match simulation with data as well as possible & find the right mix
• Used 500 ADC thresholds in CLC
• Used 3 GeV threshold in plug• East and West coincidence in
both• CLC+ plug acceptance in MC =
94%
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CLC Min. Bias Trigger
•Min. bias trigger requires at least 1 hit from East AND West CLC modules
•Loose time gate used to reduce losses contributions
•CLC min. bias trigger efficiency is studied with using orthogonal triggers to tag min. bias events
•L1_CEM0_PT4, L1_CMU0_PT8
•L1_CEM4_PT4, L1_CEM8_PT8
•L1_CMU1.5_PT1.5, L1_CMU6_PT8
• =
• =hctrig meas
tagmeastag Ntrig*tag/Ntag
/(1 - flosstag
- fnon-emptytag
)
Efficiency for tagging on hardcore events (corrected for losses and non-collision events, and assuming negligible diffractive
contributions in tagged events)
Roberto RossinRoberto RossinBologna CLC seminar, Bologna CLC seminar,
03/14/200603/14/2006
CLC CLC Min. Bias Trigger
•Min. bias trigger efficiency determined from samples of events passing Level-1 muon triggers
•Min. bias trigger efficiency as function of the sum of track Pt in the central region (-1< <1)
•Simulated trigger efficiency (req tag in CCAL > 10 GeV) is ~98%
~97.5%
~97%
0-Bias Trigger
L1_CEM0_PT4
0.9
1.0
0.5
Run #
Pt (GeV)