SiPM Misure

1
A new tile calorimeter with Silicon
Photomultipliers for the KLOE-2 experiment
Ivano Sarra University of Tor Vergata
Laboratori Nazionali di Frascati
Young Researcher Program _at_ Frascati Spring
School 2008 LNF- Frascati ( 13-5-2008)
2
Summary of the existing QCAL
Outline

• The proposal of a new quadrupole calorimeter
  QCALT
• A new kind of device the SIPM
• Test on SiPM (Hamamatsu MPPC)
• Test on different fiber types
• Tests on Tiles
• Conclusions

3
Summary of the existing QCAL
Summary of the existing QCAL
At KLOE the measurement of direct CP violation
is possible through the double ratio R
G(KL ?pp?) G(KS ?p0p0) / G(KS ?pp?)
G(KL?p0p0)
For the neutral decay of KL ? 2p0 ? 4?
To recover photons lost on the quadrupole region
the area is covered by a Tile Calorimeter QCAL
4
Proposal of new QCAL
Proposal of new QCAL
For the high precision measurement of KL?2?0
decay rate - Adapt a new calorimeter over new
interaction region - Improve granularity, time
resolution efficiency.

• Barrel with 12 modules
• - Each module has a thickness
• of 5-6 cm and ?1 m length.
• It is made by 8 layers of
• 2 mm W /3 mm Scint.

Z
R
Along Z, each slab is divided in 20 tiles of 5x5
cm2 Tile dimension increases along R.
Z
5
New tiles design
New tile design
The RD for Tesla/ILC made possible a very
promising tile detector - Square tiles with
fibers in circular grooves. - Tile readout is
possible with SiPM
SIPM SILICON PHOTOMULTIPLIER Array of Single
Geiger Mode APD. It is a discrete detector for
photon counting depending on the PIXEL size
MPPC SIPM by Hamamatsu 1 mm2 area 100 pixels
--gt 100 um 400 pixels --gt 50 um
6
Test on SiPM
7
First study on SiPM
First study on SiPM

• To study SiPM characteristics we use
• Black box
• Pulsed led to fire SiPM
• Polaroid filter to change light intensity
• We can measure
• Gain vs Vbias
• Gain vs Temperature
• Dark noise rate

8
SIPM signal with BLUE Led Pulser
From Scope
Vbias 69.25Volt, T24C Rise Time 3ns, Fall
Time 150ns
From Adc
From ADC spectra, we get single photoelectron
charge (Vbias 69.25, T24C) Q 0.36pC Gain
2.3E06
?count17.4 Q17.40.25pC4.35pC Q4.35/11.8(ampl
.)0.36pC GQ/e
9
Gain vs T
Our result
Vbias69.30V
Hamamatsu
?G-0.12 ?T
10
Dark Count vs Vbias
Dark-Count(kHz)
Our result
Hamamatsu
11
Test on fibers
12
Test of single Scintillating Fibers

• We have studied the characteristics of 3
  different types of fibers
• Kuraray SCSF 81 (Blue )
• Saint Gobain BCF92 single cladding (Green)
• Saint Gobain BCF92 multi cladding (Green)
• The test is performed using SiPM and a beta
  source of Sr90.
• The trigger is provided by a NE110 finger (1cm x
  5cm) readout by 1 PM.

13
Selected Scintillating Fibers

• After the test we
• have selected
• Saint-Gobain Multi
• Cladding fibers
• 1) Best light yield
• 2) Fast emission time
• (3-4 ns/p.e.)
• 3) High attenuation
• length (3.5 m)

Q( ADC COUNTS)
14
Test on tiles
15
Test of tiles

• 3 possible solutions under study
• 1) SIPM directly on tile
• 2) SIPM amplifier HV on tile
• 3) SIPM connected to fibers in a
• far-away position from tile
• At the moment we have tested
• only the third solution
• Tiles 3mm and 5 mm thickness
• Without reflector at fiber end
• Simple mylar around tile
• SiPM placed outside tile in
• optical contact (w grease)

16
Test of Tiles

• Data taking with cosmic rays.
• Trigger using 2 scintillator counters read at
  both ends.

• Tested 2 tiles with different thickness and
  different SIPM.
• To investigate the use of SIPM_at_400 pixels (vs
  SIPM_at_100 pixels) which has
• a gain reduction of 1/3 (7.5 105 instead
  2.4E106)
• a reduced temperature dependence DG -0.03DT
  (instead -0.12)

17
Test of Tiles (MIP distribution)
ADC distributions for two different
thicknesses The MIP values are compatible
taking into account different thicknesses and
QE of the two SIPMs. N3mm N5mm
x 3/5 x 0.40/0.45 N3mm 14
3mm thick 400 Pixels SIPM ltMIPgt 14 pe
5mm thick 100 Pixels SIPM ltMIPgt 26 pe
18
Tile test (time resolution for MIP)
110 ps/counts
5 mm thick
3 mm thick
TDC ( Counts)

• After correcting the pulse height dependence on
  the timing,
• a Time Resolution of 750 (1000) ps is obtained
  for a MIP
• on the 5 (3) mm thick tiles.
• No correction applied to the trigger jitter.

19
Conclusions and plans
SiPM our tests confirm Hamamatsu characteristics
for 100 pixels MPPC - Gain vs HV - Gain vs
temperature - Dark noise Reduced temperature
variation of gain and dark noise expected for a
400 pixels MPPC (50 ?m pixel). Fibers adopted
solution is the Saint Gobain multi
cladding. Tile Good results on light response
and timing. Light yield and time
resolution sufficient for our purposes.
Solution with MPPCamp directly on tile under
development.
20
Spares
21
Set Up

• - HV stability 10 mV
• - Blue LED diode on
• SiPM
• Temperature measured
• on SiPM
• CAMAC DAQ
• ADC sensitivity
• 0.25 pC/cnt

22
Mppc Multi Pixel Foton Counter

• Mppc Multi Pixel Foton Counter -100C N.370,
  characteristics at 25C and ?655 nmVop.
  69,28V, Gain 2.41E6

23
The KLOE experiment

• The KLOE design was driven by the measurement of
  direct CP violation
• through the double ratio R G(KL ?pp?) G(KS
  ?p0p0) / G(KS ?pp?) G(KL?p0p0)
• Collision at sqrt(s)Mphi 1.02GeV
• (e-e)? F ? (kS kL) (k- k)

• Electromagnetic Calorimeter
• Measure charged particles
• lead/scint. fibers
• 4880 PM

• Drift Chamber
• Measure charged particles
• (4 m thick ? 3.4 m lenght)
• ? 90 He 10 iC4H10
• 52140 wires

Superconducting coil B6kGauss
24
Dark Count shape vs Vbias
T 24 C

• VRIRQ/t,
• Where
• t 35ns
• R 50O
• Dark rate follows
• specifications.
• It becomes negligible
• when triggering at
• 1.5 pe.

25
Tile test
Time resolution measured using different number
of photoelectrons on tile.
Result compatible with 5mm tile. No trigger
jitter corrected. Stochastic term roughly
consistent with
26
Fibers test
Saint Gobain multi cladding
0pe
1pe

• Pedestal
• Cut _at_ 0.5 pe
• Cut _at_ 1.5 pe

2pe
3pe
4pe
5pe
27
Fibers test
Saint Gobain single cladding
1pe
0pe

• Pedestal
• Cut _at_ 0.5 pe
• Cut _at_ 1.5 pe

2pe
3pe
4pe
28
Fibers test
Kuraray Y11
0pe
1pe

• Pedestal
• Cut _at_ 0.5 pe
• Cut _at_ 1.5 pe

2pe
3pe
29
Tile test
Entries
ADC distribution obtained using a 3mm tile
optically coupled with a 400 pixels SiPM.
0pe
1pe
ADC counts
30
Tile test
Using 3mm tile with 400 pixels MPPC. Slewing
correction. Fit function
TDC Vs ADC
Charge of imput signal ADC counts
31
Apd operanti in Geiger Mode

• Diodo a Vbias gt Vbd
• t lt t0 ... i0, non ci sono portatori
• t t0, inizia la valanga
• t0 lt t lt t1, la valanga si diffonde
• t gt t1, la valanga si auto-sostiene ed è
  limitata ad Imax dalle resistenze in serie

Meccanismo di Quencing
32
Apd operanti in Geiger Mode
Gli Apd operanti in geiger mode possono essere
modellati tramite il seguente circuito
elettronico

• Switch Open quando la valanga non è innescata
  Cd si carica a Vbias e non scorre corrente
• Switch Close quando la valanga si innesca Cd si
  scarica fino a Vbd con tRsCd e la corrente va
  ad I(Vbias-Vbd)/RQ

tQRQCd35ns
33
Gain vs Vbias.2
From Hamamatsu

• Our measurement

34
Gain vs Vbias
ADC spectra as a function of the applied HV.
Gate 350ns T24C
35
Gain vs Vbias
Increasing HV we increase dark rate
Vbias 68.94V
Vbias 68.99V
Vbias 69.05V
36
Gain vs Vbias
?G2.24 ?V ?G2.19?V ?G2.12?V
Our result
Hamamatsu
?G2.25 ?V