The HERA-B Ring Imaging
Cherenkov System –
Design and Performance
Beauty ‘99
Bled, Slovenia, June 24, 1999
Jörg Pyrlik
University of Houston, HERA-B Collaboration
this RICH is a
joint effort of groups
from
University of Texas at Austin
University of Barcelona
University of Coimbra
DESY
University of Hamburg
University of Houston
University of Ljubljana & J. Stefan Institute
Northwestern University
Jörg Pyrlik, Beauty 99
1
Content
❏ Design
■
RICH in HERA-B
■
Radiator vessel and mirrors
■
Focal plane layout
■
Photon detector optics, mechanics, electronics
❏ Running
■
Commissioning milestones
■
Occupancies, distributions and events
❏ Performance
■
Reconstruction software
■
Rings (Number of photons)
■
Resolution
■
First particle identification
❏ Conclusion
Jörg Pyrlik, Beauty 99
2
RICH in HERA-B
❏ HERA-B
■
Fixed target experiment at the 920 GeV HERA
proton ring at DESY
■
Goal: measure CP violation in B-system
■
Hundreds of charged tracks every 96 ns
❏ RICH
■
Particle identification; tag flavor of B decays
3σ π/K separation for 10 < p < 75 GeV/c
■
full acceptance: ±160 mrad vertical,
±250 mrad horizontal (bending plane)
➤ Single RICH in the middle of the detector
➤ Photon detectors outside of particle flux
4m
0
RICH
20
Jörg Pyrlik, Beauty 99
15
10
5m
0
3
RICH at Large
Photon
Detectors
❏ Radiator
■
100 m3 stainless steel vessel
■
C4F10 n = 1.00135
■
+160 m
rad
Θ‰erenkov = 52 mrad
❏ Mirrors
■
Spherical (r = 11.4 m),
tilted up/down, 80 hexagon
based elements
■
Planar mirrors move focal
plane above and below
flux of particles
11.4 m
rad
-160 m
C4 F10
Spherical
Mirrors
Photon
Detectors
Planar
Mirrors
4m
6m
Jörg Pyrlik, Beauty 99
4
Photon Detector
❏ Focal Plane
■
Best approximated by
two cylinders
■
Light exits through
2 mm thick UV grade
Plexiglas windows
■
Incidence angle ≈ 2 θC
due to unknown z of photon emission and
effect of magnetic field
❏ Multi-Anode Photomultipliers
■
Hamamatsu R5900 with
18 × 18 mm2 photocathode
■
1500 with 16 anodes
750 with 4 anodes
M16
M4
❏ Mounting
■
Soft iron box (supermodule) with 32×12 mounting positions
■
PMTs placed
according to
occupancy
empty
Supermodules in upper focal plane
Jörg Pyrlik, Beauty 99
5
Photon Detector Parts
❏ Lens Module
■
2-lens telescope, maps
36 mm ¨ onto 18 mm
¨ photocathode
■
Optimized aspheric
lenses, molded from
UV-acrylate
■
Holder with 12 telescopes (2 × 6) placed
on supermodule
❏ Supermodule
■
Holds lenses and
PMTs/electronics
■
Provides magnetic
shielding
❏ Electronics
Jörg Pyrlik, Beauty 99
■
Passive base-board
with socket for 4 PMTs
(M16 or M4)
■
16ch amplifier-shaperdiscriminator cards
plugged in on back
6
Commissioning
❏ Radiator Vessel
■
Completed in February 1998
■
Mirrors installed and aligned April 1998
■
First C4F10 fill end of December 1998
❏ Photon Detector
■
1488 M16 and 752 M4 PMTs (including frontend electronics) installed in April 1998
■
Lens system completed in May 1998
■
Readout chain for all 27,000 channels complete
in Summer 1998
❏ Physics
Jörg Pyrlik, Beauty 99
■
“First Light” with
air as radiator on
August 19, 1998
■
First “Big Rings”
using C4F 10 in
January 1999
■
Magnet-On data since May 1999
7
Performance
❏ Occupancy, C4F10, Magnet OFF
■
Jörg Pyrlik, Beauty 99
System works as expected:
▼
Low angle events populate center
▼
M4s have four times more hits than M16s
▼
Ring shared between upper and lower half
▼
Few dead R/O cards and PMTs
▼
Very few hot channels
8
Performance
❏ Occupancy, C4F10, Magnet ON
■
Jörg Pyrlik, Beauty 99
The magnet really works:
▼
Field separates charges
▼
More even distribution of occupancy
▼
Same dead R/O cards and PMTs
▼
Most dead channels have been already fixed during
recent shut-down
9
Distributions
❏ Occupancy Comparison
■
Magnet removes low-momentum junk
■
Mean occupancy (including empty events)
■
▼
magnet ON
422
▼
magnet OFF
1155
Triggered data bias towards events with high
multiplicity
Zero suppressed
Hits per event (triggered data)
30
20
Magnet ON
Magnet OFF
10
0
0
Jörg Pyrlik, Beauty 99
1000
2000
3000
4000
5000
10
Events
❏ All Magnet On
❏ “Typical” (754 hits)
EVENT: 205 4897
Jörg Pyrlik, Beauty 99
Apr 13 14:42:15 1999
11
Events
❏ Good (172 hits)
EVENT: 1111 4897
Jörg Pyrlik, Beauty 99
Apr 13 14:42:42 1999
12
Events
❏ Bad (2184 hits)
EVENT: 335 4897
Jörg Pyrlik, Beauty 99
Tue Apr 13 14:42:21 1999
13
Software
❏ Ring Reconstruction
Declination λ [mrad]
Detector mapped into λ -ϕ space
-100
0
100
■
Azimuth ϕ [mrad]
-20 0
■
-10 0
0
100
200
Stand-alone ring search algorithm
▼
Histogramming method
▼
Ring of correct radius
gives peak at center
■
Correct for distortions
■
Refit „ellipse“ using Gaussian
➤ Implemented into standard
HERA-B analysis
Jörg Pyrlik, Beauty 99
14
Reconstructed Rings
❏ Software finds rings
arte vs= 7.0100
prism 4.11/04b
EvtRunExp = 2646 5429 1
λ
-50
0
-50
-100
-100
-50
0
50
λ
100
ϕ
arte vs= 7.0100
prism 4.11/04b
EvtRunExp = 2646 5429 1
-50
0
-50
-100
-100
Jörg Pyrlik, Beauty 99
-50
0
50
100
ϕ
15
Predicted No. of Photons
❏ Simple Model
▼
▼
6
Generation of ‰erenkov
light in pure C4F10
Reflectivity of mirrors
5
4
3
▼
Window transmission
▼
PM quantum efficiency
▼
Telescope efficiency
Photon Enery [eV]
2
50
Generated Photons
40
Photons / 10nm
Photons reaching
exit window
30
Photons
exiting vessel
20
PMT
QE
10
Detected photons
0
200
300
400
500
Wavelength [nm]
Prediction for β = 1: 34 ± 2 photons / ring
Jörg Pyrlik, Beauty 99
16
Ring Analysis
❏ Photons per Ring for β = 1 particles
Magnet-off
Magnet-off data
data using
usingCalorimeter:
Calorimeter:
Select
Selectrings
ringswith
withEEclust >>88GeV
GeV
500
clust
400
Mean
Mean photons
photons //ring
ring
MC:
36.2
MC:
36.2
Data:
35.4
Data:
35.4
300
200
100
0
0
20
40
60
Nphotons
80
❏ Single hit resolution - Sigma of ring
Magnet-off
Magnet-off data
data using
usingCalorimeter:
Calorimeter:
Select
Selectrings
ringswith
withEEclust >>88GeV
GeV
400
clust
300
200
Expectation:
Expectation:
dispersion:
dispersion:
granularity:
granularity:
0.4
0.4 mrad
mrad
0.5
0.5 mrad
mrad
Data:
Data:
0.65
0.65 mrad
mrad
100
0
0
Jörg Pyrlik, Beauty 99
2
4
6
8
Very good
Agreement !
σ [mrad]
17
Ring Variations
❏ Radiator Gas System
■
C4F10 „closed“ recirculation system
■
Radiator at 1.5 mbar above atmospheric
pressure
➤ Ring radius changes with weather
Ring radius [mrad]
1020
1010
1000
990
mbar
52.4
52.2
52
51.8
51.6
Barometric Pressure
51.4
31.5
31.6
31.7
31.8
31.9
√mbar
➤ We have some leaks !
N2 concentration [%]
16
12
8
4
0
Jörg Pyrlik, Beauty 99
20
40
60
80
Days in 99
18
Particle ID
❏ Main task of RICH: Particle Identification
■
No tracking yet, magnet on only recently
❏ Determine momentum with RICH and ECAL
■
ECAL space point: Eclust > 4 GeV
■
Track direction: RICH ring
■
Match ring and cluster:
| by | < 2 cm
ECAL
RICH
Impact b
x
parameter
Magnet
Track
direction
Target
Space
point
p0 = ∫ Bdl 3.3 T ⋅ m GeV
Momentum:
Resolution:
Jörg Pyrlik, Beauty 99
bx
bx
1
=
≈
p p0 zbend 280 cm ⋅ GeV
δp
p
≈ 0.03 ⊕
p
280 GeV/c
19
Particle ID
❏ RICH-ECAL measured quantities
■
‰erenkov relation, cosθC = 1 ⁄ n β, relates the
angle to momentum:
θ C2
■
= θ 02
m2
− 2
p
θ0 ≡ “β = 1”
θC2 vs. 1 ⁄ p2: particles of same mass on
straight lines
2.5
1000 θ C2
2.0
1.5
1.0
e
π
0.5
K
p
0
80
Jörg Pyrlik, Beauty 99
30
20
15
p [GeV]
10
20
Conclusion
❏ RICH works as advertised
❏ Powerful Detector with great potential
❏ We are eager to see the rest of HERA-B
being completed
Jörg Pyrlik, Beauty 99
21