KTeV Detector and Decay 0 00

1
KTeV Detector and Decay ?0 ? ?0?0?
University of Wisconsin_Madison
Huican Ping 08/19/2003
Contents KTeV detector ?0 ? ?0?0?
analysis Future Plan
2
Overview of KTeV apparatus
1
1 Beamline Part Beam, Target, Sweeping
Magnets, Absorbers, Collimators, Spin Rotator
Magnet 2 Decay Region Pressure, Size,
RCs. 3 Spectrometers Four drift chambers
(DCs), Analyzing Magnet 4 The Photon
Veto System Ring Counters, Spectrometer Antis
(SA2, SA3, SA4), CsI Anti (CIA), Collar
Anti. 5 The Transition Radiation Detector
Radiator, MWPC, Gas, Performance 6 Cesium
Iodide Calorimeter (EM Calorimeter) CsI
Crystals, PMT, Readout Electronics,
Performance 7 VV Trigger Hodoscopes, Hole
Counter, The Muon System 8 Trigger System
Level1, Level2 and Level3 Triggers
3
Beamline
2
Primary Beam
Primary Beam 800 GeV primary proton beam
hits the BeO target. The secondary beam consists
mainly of neutral kaons, but includes others
charged and neutral particles. Protons arrive
in pulse 1-2 ns wide in a RF-Bucket of 19ns
(53MHz frequency). A spill is 23 seconds,
off-spill periods is 37 seconds. The Tevatron
provides up to 51012 protons in every spill.
The BeO rod is about 33 mm across and 30 cm
long. The target is cooled by a flow of nitrogen
to keep temperature about 125 0C.
Why BeO as target We notice that for high Z
targets, there is a very large photon flux
produced, a low Z target is favored, so BeO. BeO
has higher melting point than Be.
From FermiLab KTeV Conceptual Design
Report June 7, 1991.
GEANT Particle Yields Normalized to
1013 protons on a 5 cm target and 3
microsteradians neutral beam. All entries107
4
Beamline
3
Collimators is to define the beams in the
transverse direction.
The Sweepers to deflect charged particles There
are 3 sweepers. The Field is in y-axis. Basically
the Target Sweeper is to deflect the remains of
primary protons, Sweeper1 is to remove muons,
Sweeper2 is to remove interaction Charged
particles. Absorber(Pb) to get rid of some
neutral particles, e.g. Photons (Pb).
Spin Rotator Magnet The 3 Sweepers are tuned to
rotate the ? and ?s polarization vector from
x-axis to z-axis beam direction. Spin Rotator
Magnet is in x-axis, rotates the polarization
vector from z-axis to y-axis. Its dipole are
reverse every half day. The spin rotator had no
effect on the Kaons.
?0 spin Proton Direction p z 4.810-3y ?0
Direction ?0 ? z Parity conserved, ?0 spin
direction p ? ?0 Spin along x axis, ? to
Production Plane
Finally Neutron (3.5), KL(1.0) and high momentum
?(0.02) and ?(0.002) can go to the following
decay region.
5
Decay Region
4
General The tank is 65 meters long, which
is from Z93.0m to Z158.9m. It starts as a
45.73 cm diameter to a size of 243.84 cm in
diameter. Pressure is 10-6 torr to reduce the
interactions between the neutral beams and any
surrounding matter, and to reduce scattering of
charged decay products.
6
Spectrometer (1)
5
By multiple-scattering
4 Drift Chambers, Analyzing Magnet
Resolution s (p) /p 0 .38 ? 0 .016p
Momenta, Vertex of Charged Particles
Finite resolution of measured hit position 100
um
General Each DC is made up of two views
of planes. The two planes in each view are offset
by one half-cell(6.35 mm) to permit an
unambiguous hit position. Wire Sense wires
and Field wires. Sense wire is 1 mil gold-plated
tungsten and field wires are 4 mil gold-plated
aluminum.
Schematic Drawing of Drift Chamber
7
Spectrometer (2)
6
Voltage, Drift Velocity and Drift time Voltage
depends on different run, typically it is 2450
-2500V. Under these voltage, the drift speed is
nearly constant over most of the drift cell about
50 um/ns, and a drift time of less than 200 ns
for electrons. The electrical pulse on the sense
wire is amplified in two stages and discriminated
with the time measure by a TDC with a precision
of 0.5 ns.
Gas Half-half of argon and ethane, along with
1 isopropyl alcohol. Alcohol absorbs ultraviolet
light during gas amplification and as quencher,
but it causes deposits on wires.
Sum-Of-Distance Hits from good tracks should
reconstruct to a Sum-Of-Distance(SOD) 6.35 mm
of the first electron.
Analyzing Magnet Located at Z170m, Dipole
magnet (3.05m in Z axis, 2.90m in X axis).
Vertical field 2000 Gauss, Momentum kick
0.205 Gev/c. The field orientation is reversed
once per day in order to remove systematic
biases relating to the magnet polarity.
8
Momentum Resolution
By multiple-scattering
7
KTeV Resolution s (p) /p 0 .38 ? 0 .016p
Finite resolution of measured hit position 100
um
?0 (13.6MeV/?cP) Sqrt(x/X0)10.0038?ln(x/X0) d
P/P d?/? dP/P ? (13.6MeV/?cP) Sqrt(x/X0) /
? dP/P ? (13.6MeV/?c)Sqrt(x /X0)/(P?)
?x1 ?x2 ?x (??)2 (?x1/b)2 (?x2/b)2 (??)2
2(?x/b)2
mv2/R evB P eBR ? ? L / R P eLB / ? dP
eLB d? / ? 2 dP/P (eLB d? / ?2) / (eLB /
?) dP/P d?/? (2?x/b) / (eLB/P) dP/P
(2?x/b) / (eLB)P
? ?f ?r (??)2 (??f)2 (??r)2 (??)2
4(?x/b)2 ?? 2?x/b
9
Photon Veto System

To veto the events where a decay
product leaves the fiducial volume of the
detector.
8
Spectrometer Antis
Ring Counters
RCs Ring Counters, SAs Spectrometer
Antis, CIA Cesium Iodide Anti,
CA Collar Anti,
Collar Antis
10
Why 0.5 X0
9
Why people prefer to 0.5-X0/sheet Pb than that of
X0?
In order to get the efficiency for low energy
photons, it is necessary to have rather fine
sampling. 0.5X0 sampling provides reasonable
efficiency at 100MeV, it is dramatically better
than 1X0 sampling. This provides for good
efficiency for low energy photons while providing
adequate depth for detecting higher energy
photons with very high efficiency.
11
Transition Radiation Detector
10
Purpose Efficient ?-/e- rejection Level-2
triggers
General Stuff Immediately after DC4. The
TRDs account for roughly 11.5 of X0 of material
in the acceptance region (8 TRD modules). Each
TRD consisted of a radiator, active MWPC volume,
and Buffered gas volumes.
Gas
Xenon was a main
component of active volume gas, gas mixture of
80 Xe and 20 CO2 was chosen. The typical high
voltage of 2250V with a drift velocity of 5
cm/us.
The buffer volume has gas of C2F6.
Top view
Overall Single Track ? Rejection Capability
Better than 2761 rejection factor of charged
pions at 90 electron acceptance.
12
Cesium Iodide Calorimeter (1)
11
3 Purposes To measure the energy and position
of photon Particle identification like e/?
(better than 5001 for 90 electron acceptance)
Measuring the charged particle position at the
calorimeter enabled us to resolve the x-y
ambiguity in the track reconstruction.
Why CsI Crystals? Instead of Lead-glass, CsI
crystals is with good radiation hardness (CsI is
much more resistant to radiation damage comparing
to the Pb-glass), excellent energy and position
resolution, quick signals (decay time 20ns with
315nm for fast peak) and high efficiency, etc..
13
Cesium Iodide Calorimeter (2)
12
DE/E

• Geometry of CsI
• 3100 pure CsI crystals with a length of 50cm
  (27X0).
• 2232 inner crystals size 2.5x2.5 cm
• 868 outer ones 5.0x5.0 cm
• Each crystal is wrapped in 13um thick mylar, part
  of which is aluminized. By choosing where we put
  the reflective mylar, we can tune the light
  collection to be uniform to within 5.
• Since it is 27X0 long, the E deposited by EM
  interaction, such as photons and electrons, is
  almost entirely contained within the calorimeter.
  However, it is only 1.4 hadronic interaction
  lengths long, showers due to hadronic backgrounds
  such as ?- mostly go through the calorimeter .

s(E )/E 0 .45 ? 2/vE,
Position Resolution 1mm
E resolution
is better than 1 for energies above 4 Gev, and
e/? rejection is about 5001
The vE from photo-statistics in scintillation
light, The 0.45 is from light leakage,
fluctuation of light yield due to T variation
(-1.5 per degree Celsius), electronic noise ...
14
CsI Calorimeter Readout
13
PMTs with gain 5000, voltage 1200V. The
digital photomultiplier tube base (DPMT) was used
for readout. The main components of the DPMT
were a charge integrating and encoding(QIE)
custom integrated circuit, an Analog Devices
AD9002 8-bit flash ADC (FADC), and a
driver/buffer/clocking(DBC) custom integrated
circuit. The system was operated at 53MHz.
Bias Circuits
Sync bit
Cable Drivers
Sync 1
QIE
Current source (PMT)
Exponent 4
Exponent (4bits)
Analog
FADC 8 bit
Mantissa (8bits)
Clock
Digitizer Circuit Block Diagram
From KTeV0203 by R.J. Yarema, G.W.Foster etc
The QIE Divides the total current into nine
binary range(I/2, I/4, ... , I/512). The QIE
integrates binary weighted divisions of the input
signal for a period of time, selects the
integrated signal in the range of interest and
output in digital form as a 4-bit exponent, also
presents it to an 8-bit FADC. The FADC will give
the 8-bit mantissa.
15
VV Trigger Hodoscopes
14
Purpose To count the charged particles at the
trigger level1 on a much faster timescale (around
15ns) than the drift chambers could provide.
It is located before CsI calorimeter. The V V
planes are offset by one-half counter, for
reducing the particles slipping through the Crack.
Beam goes into page
Hole Counter
Why HC? A typical hyperon decay produced a high
momentum proton (anti proton) which stayed in the
beamline. HCs are used to trigger such events.
Muon System(Mu2 MU3)
Components Hodoscopes reduce the trigger rate
by rejecting events decay to muons. Steels Beam
Dump.
16
Level 1Trigger
15
Sources 1. Signals from scintillator
counters 2. Calorimeter total energy sum, et 3.
Logical OR of DC signals Rate is from 50kHz.
Takes lt 19 nsec to make decision.
Purpose of Logical of DC signals In order to
reduce the level 1 rate, the DCOR were designed
and built to quickly detect drift chamber hits
present in an event.
How Each group of (14)16 wires (from
twisted-flat ribbon pairs) of each plane pair is
connected to a single DCOR module. It stretched
the 40 nsec drift chamber pulse to 90 nsec, about
10 nsec greater than half the nominal maximum
drift time. Outputs from the X-X(Y-Y) planes
were added in groups to form effective hodoscope
paddles for use in the trigger logic at NIM
level.
17
Level 2Trigger
16
Sources 1. DC hit counting (It takes around
800 ns to provide a result), 2. CsI Hardware
Cluster Counting (The HCC has the longest
decision time of all the Level 2 processors about
1.5us) 3. TRD information. The L2 rate is
about 10kHz. The trigger took a longer time,
about 3 usec. Prescale is used if the event rate
is too high before the Level 2 trigger. If an
event failed at L2, then aborted before finishing
ADC digitization, to next event.
HCC -- Algorithm
Basic Idea A cluster has a perimeter which
completely encloses the cluster. A perimeter has
a net total of 4 convex turns. Bit-wise
operation for each crystal.
18
The Stiff Track Trigger (STT)
17
Important for hyperon decay in KTeV Hyperon
decay modes have a high momentum proton with the
beam axis.
Logical We calculate the quantity
(xDC4-xDC3)-(xDC2-xDC1). This is proportional to
the change in slope of the track. If its value
was between 7 and 7 inclusive, the event was
selected by STT. Any other combination of inputs
caused the STT to reject the event.
L1 and L2 Hyperon Trigger
Three bits (10th, 11th,12th bit) of the beam
trigger were dedicated to the hyperon data.
19
Level 3 Trigger
18
For example KL ? ??- trigger Tracks and
Vertices candidates are reconstructed The
invariant mass is calculated for each vertex
candidate. If any vertex candidate has an
invariant mass of over 450MeV, the calorimeter
information is unpacked and clusters are found
and matched to the tracks. The quantity E/p is
calculated for each track and required to be less
than 0.9 in order for the event to be tagged as a
??- candidate. Then Candidate ??- are tagged
and be written to tape. In all, about 40,000
events are written to tape from each accelerator
spill at nominal intensity.
Sources Software filter code running on 24
multiprocessing Challenge SGI CPUs (200MHz). The
Rate is 2 KHz, it takes 3 ms to process one
event, Event size 8 KB. All digitization was
completed before the Level 3. L3 made a final
decision of writing it out to tape devices, based
on the selection criteria imposed on the tagged
trigger.
3msec
3msec
0
L1 trigger
L2 trigger
L3 trigger
20
Part II.
Analysis of Normalization mode ?0 ? ?0
?0 And ?0 ? ?0 ?0 ?
21
Why ?0 ? ?0?0?
20

• ?0 ? ?0 ?0 ? decay is a neutral weak radiative
  decay.
• There is no detailed theoretical or experimental
  literature reported.
• ?0 Decay Mode
• ?0 ? ?0 ?0 (99.522 ? 0.032)
• ?0 ? ?0 ? (1.18 ? 0. 30) o
  else
• ?0 ? ?0 ?0 ? not
  available up to now.
• KTeV experiment provides enough raw data about
  Cascade decay modes. For example, ?0 ? ? e- ?e
  , ?0 ? ?0 ? etc were studied.

22
?0 ? ?0 ?0 Selection Cuts
21
In general, events with equal and more than 2
neutral clusters were selected. The 3 highest
energy neutral clusters were combined into 3
possible pi0 pairs. At the end, if after the
application of the above cuts more than one ?0
candidates were found, the one yielding the
lowest pt was kept.
23
Preliminary ?0 ? ?0 ?0 results
22
Monte Carlo Data Generated Monte Carlo
3,999,908 events Remained after Cuts
107,316 events Detection Efficiency
(2.683?0.008)x10-2
DST Tape Data DST Tape
Events Efficiency corrected
UPH001 UPH060 (99) 2,720,683 ?
1649 101,404,510 ? 308,535 KQHY042KQHY046
(97) 274,299 ? 524 10,223,592
? 36,197 TOTAL
2,994,982 ? 1731 111,628,103 ? 338,824
There are about 110 million events in 65 DST
tapes.
24
?0 ? ?0 ?0 MC- Real Data Comparison (1)
23
25
?0 ? ?0 ?0 MC- Real Data Comparison (2)
24
26
?0 ? ?0 ?0 ? Selection Cuts
25
27
Events Re-construction
26
28
Some Important Cuts
27
29
?0 ? ?0 ?0 ? Branch Ratio (Preliminary)
28
30
?0 ? ?0 ?0 ? MC - Real Data Comparison
29
31
?0 ? ?0 ?0 ? Background Study
30
We studied several possible backgrounds
Should study more possible backgrounds in the
future.
32
Conclusion and Future Plan
31

• It is hopeful that there are signals of the
  decay mode
• ?0 ? ?0 ?0 ? with the preliminary BR( 5.048?
  1.106) ?10-5
• A lot of work is needed
• More Careful and Systematic Study/Analysis.
• Other background sources?
• Theoretical prediction?
• Troubling
• 1 Width of ?0 mass
• 2 Pt2 distribution of ?0 ? ?0 ?0 ?
  events is not good.

33
Thanks!