Development of scintillation detectors with MRS APD light

Development of scintillation detectors with MRS APD light readout for CBM Muon system and ECAL preshower Alexander Akindinov ITEP(Moscow) on behalf of ALICE-ITEP group

Outlook • MRS APD – fast Geiger mode high gain photodetector • START – technology • ALICE CRTF – first implementation for START technology • Counters for CBM muon systems • Counters for EMCAL preshower • Conclusion

MRS APD – fast Geiger mode high gain photodetector M = C*(U - Ubr) MRS APD structure Readout -Voper Many other name: Si. PM, MPPC, GAPD… Slope corresponds to elementary micro-cell capacity Q/ U. Q = Ncell*C*(U - Ubr) Single electron resolution is a result of dispersion of Ubr and Ccell across diode area.

Optical trench(CPTA) – low optical coupling between cells. Acceptable noise rate for reasonable thresholds Noise spectrum slope is a result of “red” cross-talks between cells, when avalanche in one cell initiate avalanche in neighboring one via photon. Probability to have more than one pixel fired from a single primary e-h pair = (N n+1)/ (N n ) PDE is optimize for WLS fibers P. D. E. = Q. E. * g* R εg – geometrical fill-factor R – probability to initiate Geiger avalanche Q. E. – quantum efficiency CPTA MRS APD α= 0. 1 For 4 th cell threshold noise count about a few k. Hz High geometrical factor

START- (Scintillation Tile with MRS APD Light Readou. T) MRS APD optimize geometry (d=1. 15 mm) for the fiber readout σ-shape groove for uniform response Two MRS APD in coincidence “On board” FEE Mosaic array Threshold Signal/noise Cosmic μ Polisterol 165 150 X 10 mm 1 groove, 2 rings Y 11 and glued

First implementation - ALICE CRTF Upper plane 10 strips 8 STARTs in each (80 STARTS) Bottom plane 10 strips 8 STARTs in each (80 STARTS) Up to 5 ALICE TOF modules under the test.

Trigger and monitoring Simple logic block for trigger configuration and monitoring were connected via USB to Lab. View • !!Monitoring the only by counting rate of individual detector, no any LED system was used. !!! • Efficiency was measured as difference in counting rate for 80 pairs.

Long term operation results First real experiment with 320 new photosensors (CPTA firm MRS APD). About 3 years of continues operation, all 87 ALICE TOF modules was tested. Summary of “dead” and replaced diodes Reason Number of diodes Explanation Action Leakage current increasing up to 20 μA 27 Bad insulation between metal and cell for trench technology Informing CPTA, work in progress Zero current 4 Bad bonding Negligible, but informing CPTA Total: 31 !!!!<10 % for 3 years continuos operation!!!!!

Counters for CBM muon system INTAS Ref. Nr 06 -1000012 -8781 “Develop and produce START prototypes adapted for d. E/dx measurements. Five detectors of each of two types of START will be produced: tiles sized 100 x 10 mm 3 and strips sized 5 x 200 x 10 mm 3 Strips 5 x 200 x 10 mm 3 Tiles with different groove diameters to find optimal uniformity (d = 90, 80, 70, 60, 50 mm) ROLAND milling machine !!!Many thanks to our friends from ITEP-CALICE group!!!!

New package to minimize “dead zone” and new connector TO-18 Very uniform response (about 8%). Main sources - regions near optical connectors and groove. New transparent optical connector produced by the same digital milling machine as tiles New package

Diodes calibration and fiber cut Setup to cut fiber with microscope Each diode after placement into connector was checked (how well diode active area is with connector hole) Setup for diodes test Mixer 12 channel board

Diodes calibration and fiber cut I-U curve to find out breakdown point Number of photoelectrons for given light was checked Single cell noise and photoionization (α) coefficient were measured

Detector assembly and electronics FEE card for tiles Simple amplifier for strips (10 mm width) Tiles were wrapped in Tyveсk and light shield

FEE for tiles Ubias сontrol Test input (50 Ω) Jumper 2 for test Jumper 1 for test Ubias Amplitude 2 (50 Ω) Amplitude 1 (50 Ω) Threshold Jumper 2 for coincidence (on) Jumper 1 for coincidence (on) GND +4 -+10 V LVDS LED

First task: find optimal geometry for uniform response D=60 mm D=50 mm D=80 mm D=70 mm D=90 mm Scans were done using Ru β – source with 2. 5 mm collimator. 19 x 19 points (5 mm step). D 50 60 70 80 90 R. M. S 10. 1 8. 7 6. 6 4. 7 5. 8

New improved MRS APD (2007) Data was taken from http: //www. photonique. ch/ Photonique SA C. P. 1562 1211 Geneva-1 Switzerland 2005 – diodes for ALICE CRTF New diodes 2007 Reduced thickness of n+ - PDE increased + blue enhanced

Light yield with old and new diodes Beam test CERN PS (2 x 2 cm 2 trigger center of tile) N ≈16 ph. e. “ 43 V” old N ≈ 22 ph. e. “ 23 V”new Tile 150 x 10 mm 3 N ≈ 29 ph. e. “ 23 V”new Tile 100 x 10 mm 3 With this significant improvement can we work with one diode?

Tiles with single diodes Three reasons to produce tiles with one diode: 1. To check if we can improve light yield using another glue 2. Check if we can work with one diode 3. Make prototype for ECAL preshower 100 x 10 mm 3 with one output for fiber ( d=80 mm ) So called “half-board” designed for long strips 3 tiles were used for test 2 were assembled for cosmic test

Other task: to improve gluing How it’s important 4 grams each time Accuracy up to 0. 01 g Measurements were done before and after gluing with the same diode Does not depend on plastic type Not Bicron glued optical cement 1 1. 58 Mono-component optical glue (UV light polymerized) 1. 41 Epo-Tek 301 1. 68 Since refractive indexes almost the same we can suppose that difference is due to acid type (groove walls “melting” – more uniform and transparent volume). For our milling process – Epo-Tek 301 is the best.

Cosmic ray test one diode “Soft” Muons Ped Muons Diode noise “Soft” and “hard” part of cosmic ray It works!!! We can use only one diode.

Last task: Strips 3 M tape R. M. S ≈ 12% 4 stripes shows some “oscillating” signals. Reason bad 3 M mirror contact. All tiles were reassembled. New test will be done about 20 of October.

Conclusion • START detector was adopted for d. E/dx measurements • Optimal diameter d=80 mm for tile 100 x 10 mm 3 response was found (r. m. s < 5. 0 %). • EPO-TEK 301 glue was found as optimal to increase light yield. • 3 types of FEE board were produced (all commercially evaluable) • Progress in MRS APD development allows to use only one diode for readout keeping signal/noise ratio almost the same like for two old diodes • 4 strips 200 x 10 x 5 mm 3 were reassembled and will be tested again. Good one has 12% r. m. s. (? ) • Tools for mass production were produced.

Thank you!

BACKUP

MRS APD calibration Ubias for given amplification about 106 Thank to Yuri Musienko from CMS APD lab.

Response and leakage tests Difference – small difference in α and coupling with fiber (epoxy layer share) All current is due to intrinsic noise, no significant leakage

Noise measurements Difference – different field noise contribution to thermal noise Photoionization α