Inner HIPT Chambers Occupancies

1
Super B Factory Workshop In Hawaii 19-22 Jan
2004, Honolulu, Hawaii
Scintillator based muon upgrade / BELLE
M.Danilov, D.Liventsev, R.Mizuk, P.Pakhlov,
V.Rusinov, E.Tarkovsky, I.Tikhomirov, T.Uglov
(ITEP Moscow)
Outline

1. Scintillator strip option
2. Geiger photodiodes (SiPM MRS APD)
3. Scintillator tile option
4. First tests at BELLE
5. Conclusions

2
Scintillator Strip Geometry (1)
System consists of 17k scintillator strips with
WLS fiber readout.
3
Scintillator Strip Geometry (2)
Max. strip length 2.8m.
4
Light Yield in MINOS Scintillator Strips
1cm x 4cm, 1.2mm Y11 fiber
N p.e.
Distance along strip, m
5p.e. correspond to efficiency of about
99. Mirroring of fiber far end increases light
yield by a factor 2 (at the end).
5
Problems with PM Readout

• Complicated transportation of light with clear
  fibers.
• Difference of gain between different pixels (up
  to factor 3).
• Cross-talk 5-10 ? ghost clusters (at
  acceptable level).

6
SiPM

• SiPM main features
• Sensitive size 1x1mm2 on chip 1.5x1.5 mm2
• Gain 2?106
• Ubias50V
• Recovery time 100 ns/pixel
• Number of pixels 576-1024
• Insensitive to magnetic field
• Dynamic range 103/mm2

For further details see Advanced study of
SiPM http//www.slac.stanford.edu/pubs/icfa/fall0
1.html
7
Gain Efficiency of SiPM vs. Overvoltage
(B.Dolgoshein et al. )
Efficiency QE??geom ?geom0.3
8
SiPM Cross-talk
Interpixel Cross-talk (gt1 to 1 ratio)
Pixel Gain, 105
9
Efficiency vs. Wavelength
(B.Dolgoshein et al. )

room temperature
80
APD EGG C30626E
(NIM A428 (1999) 413-431)
60
QE,

40
PMT XP2020Q number 40979
(according to Philips Photonics)
QE

e
geom
20
SiPM
0
200
300
400
500
600
700
800
Wavelength
l
, nm
10
Emission Spectrum of Y11 WLS Fiber
Measured at distances 10cm, 30cm, 100cm and 300cm
from source.
11
SiPM Noise
random trigger
noise rate vs. threshold
1p.e.
2p.e.
Ped.
3p.e.
1p.e. noise rate 2MHz. threshold 3.5p.e.
10kHz threshold 6p.e. 1kHz
Optimization of operating voltage is subject of
RD at the moment.
12
Scan of Strip Using Cosmics Setup
groove depth 2.5mm
LED
cosmics
? Poiss?XtalkG(x0i?x,?0?1 vi)
Center of strip, N pixels (peak) 9.7
13
Scan of Strip Using Cosmics Setup (2)
Strip 200x2.5x1cm3
Poisson mean for MIP at normal incidence
MC light yield for 4x1cm2 strip will be 25
smaller.

• In the middle of 280x4x1cm3 strip we expect
  9p.e.
• With threshold 4.5p.e.
• noise rate lt10kHz
• efficiency ?98
• average efficiency gt99

Similar detector - MRS APD - with smaller x-talk
and similar efficiency is under investigation
now. Expect efficiency gt99.5 with noise lt1kHz
14
Metal-Resistor-Semiconductor APD (CPTA, Moscow)
15
MRS APD Cross-talk
16
MRS APD Efficiency vs. Wavelength
17
Scintillator Tile Geometry (1)
18
Scintillator Tile Geometry (2)
19
Light Yield from 16x16x0.5cm3 Tile with Circular
WLS
Y11 MC 1mm fiber, Vladimir Scintillator
20
Experience of CALICE (1)
21
Experience of CALICE (2)
Amplitude distributions for LED and for
triggered 90Sr signals in 5x5x0.5cm3 tile.
22
Experience of CALICE (3)
Light Yield from 129 Minical tiles.
23
Measurement of Scintillator Sensitivity to
Neutron Background at BELLE
24
Neutral KLM Clusters with Matching ECL
Originating from Background
MC (no continuum)
data
exp 13
BB bg
BB only
cos?
cos?
(lab.system)

• Random KL background exists already at present
  background rates even if coincidence between two
  layers is required.
• With scintillator KLM even one layer clusters
  could be considered for KL reconstruction.

25
Conclusions
1. Most probably RPC KLM will not be able to
work good enough at higher luminosity in end-cap
region.
2. Scintillator technique can be used to build
reliable, very efficient and simple in operation
KLM.
3. Many parameters should still be optimized.
26
Tests of Different Shapes of Grooves
27
Amplitude Dependence on Temperature
T9.5oC
T14.6oC
T20.1oC
T25.3oC
28
Light Yield from 5x4x1cm3 Scintillator with
Straight Central Groove (1mm Y11 WLS fiber,
ß-source)
29
Tests of Different WLS Fibers and Scintillator
Types
Comparison of Y11 and BC-92.
I.
To understand different light loss with BC-92 for
circular and straight grooves light loss in a
fiber loop was tested
55 cm circle
WLS
55 cm diagonal
1
1
Y11 MC 1 mm
?-source
R22mm
PM
BC-92 SC 1 mm
0.46
0.58
scint
Y11 has 6 light loss BC-92 -27 (with respect
to a straight fiber)
BICRON-408 scintillator gives x1.5 more light
than Russian scintillators. It is order of
magnitude more expensive. Therefore we use
Russian scintillator.
30
Light Yield of Individual MINOS Sc. Strips
There is a substantial difference in individual
strip responce. However all strips give more than
3p.e. at 3m.
Nonuniformity across strip is about 5 only for
4cm x 1cm strips.
31
Problems with RPC

• Hit rate at 1200mA/800mA
• Barrel 0.02 Hz/cm2 (same as no beam case)
• Endcap 0.1 Hz/cm2
• Extrapolation to x20 beam current (assuming
  linear)
• Barrel at most 0.4 Hz/cm2 at most a few
  efficiency drop, acceptable
• Endcap 2 Hz/cm2 significant efficiency drop,
  something must be done