1. Salvatore Tudisco The new generation of SPAD Single Photon Avalanche Diodes arrays I Workshop on Photon Detection - Perugia 2007 LNS LNS

2. Microelectronics The collaboration started in the 2004 to realise single device and first array prototypes - The Silicon Photomultiplier - The Silicon Photomultiplier - Arrays for imaging - Arrays for imaging

3. SPAD Si Si cathode CMOS planar technology P+P+P+P+ N-N-N-N- High Boron (P + ) concentration reduce the Breakdown voltage Buried junctions p - -p + -n - : P - high Breakdown voltage  to prevent peripheral effects P - high Breakdown voltage  to prevent peripheral effects P + to reduce the series resistance  substrate insulation  integration of many elements P + to reduce the series resistance  substrate insulation  integration of many elements P + sinkers: reduce the contact resistance of the anode and provide a low resistance path to the avalanche current N - gettering region: impurities reduction E. Sciacca et al., IEEE Trans. on el. dev. 50 (2003) 4 E. Sciacca et al., IEEE Photonics Tech. Lett. 18 (2006) 15 p-n junction reversely biased above the breakdown voltage 1  m SiO 2

4. Many elements integration  simple quenching strategy 50  100 k  Psasive Quenching SPAD Signal Rise time: 500 ps Fall time: 30 ns Recovery time: 1,5  s (increase with R L ) (increase with R L ) Important for the applications (SiPM)

5. SPAD PERPORMANCES  Photodetection Efficiency  Dark Counting Rate  Timing  Afterpulsing E. Sciacca et al., IEEE Trans. on el. dev. 50 (2003) 4 M. Belluso et al. Mem. SAIT Suppl. 9 (2006) 430 Photodetection Efficiency E.V. 10%

6. SPAD PERPORMANCES  Photodetection Efficiency  Dark Counting Rate  Timing  Afterpulsing S. Privitera et al. submited to NIMA d=20  m SPAD counter cooler

7. Laser SPAD2 TDC TRIGGER CH # SPAD1 SPAD PERPORMANCES  Photodetection Efficiency  Dark Counting Rate  Timing  Afterpulsing laser pulse 408 nm 60 ps FWHM Many photons regime   60 ps  carrier diffusion in neutral layer  delay to avalanche trigger Diffusion tail  carrier diffusion in neutral layer  delay to avalanche trigger S. Tudisco et al. Nuclear Physics B – proc. supl. 150(2006)317 Finocchiaro et al. IEEE Trans. on Nucl. Scie. 52(2005)3040 Single photon regime   120 ps

8. SPAD PERPORMANCES  Photodetection Efficiency  Dark Counting Rate  Timing  Afterpulsing Many photons regime Single photon regime laser pulse 337 nm resolution  2 ns + dye for wave length-shift

9. SPAD PERPORMANCES  Photodetection Efficiency  Dark Counting Rate  Timing  Afterpulsing S. Privitera et al. Submit. To NIMA uncorrelated dark counts Distribution of successive events to a primary avalanche Dark event After-Pulses Multi Hits TDC TRIGGER CH #

10. SPAD PERPORMANCES  Photodetection Efficiency  Dark Counting Rate  Timing  Afterpulsing Power law ? S. Privitera et al. submit. to NIMA the two contributions in 10  s the two contributions in 10  s   After the subtraction of the uncorrelated dark background

11. SPAD ARRAY: 1 st prototype 5x5 Anodes 20  m diameter, 160  m pitch Dark count rate distribution over 750 pixels (30 arrays) E. Sciacca IEEE Photonics Technology Letters 18 (13-16) (2006) 1633

12. SPAD ARRAY Optical Cross-Talk the avalanche multiplication process produce photons Isolation trench - 80  m attenuation length - 10 -5 photons per carrier crossing the junction Electrical Cross-Talk A1 A5 2mV/div 100 mV/div S. Privitera et al. submit. to NIMA 1 st observation Induction; field fluctuations  Common Substrate

13. SPAD ARRAY Time and Spatial correlations START Probability: Prompt  10 -5 Delayed  10 -3 Two contributions: Prompt < 2 ns Delayed 2 ns  3  s Pixel 5-1 Piexls 4-1, 4-2, 4-3, 3-1, 1-1 Multi Hits TDC TRIGGER CH #

14. SPAD ARRAY Time and Spatial correlations S. Privitera et al. Submit. to NIMA -Cross-talk and Afterpulsing similar trend, different slop - No correlation with distances Afterpulses 5-1 Cross-Talk STOP

15. RLRLRLRL I OUT VAVAVAVASPAD SiPM Configuration Oscilloscopio QDC or ADC channels S. Privitera et al. submit. to NIMA A.Campisi et al. NIM A 571 (1-2) (2007)350 QDCADC Laser Pulse = 408 nm = 408 nm FWHM = 50 ps Laser Pulse = 408 nm = 408 nm FWHM = 50 ps QDC channels

16. MontecarloSimulations Increasing the probability the probability Decreasing the dispersion the dispersion Event generator n° fired pixels parm. intensity Pixel ON Cross-talk generator parm. Probability Cross-talk generator parm. dispersion Fill the spectrum

17. Single device performances (20  m) PDE: ~ 45% @ 550 nm PDE: ~ 45% @ 550 nm Dark counting rate: ~ 400 cps @ 25 °C, ~ 100 cps @ 15°C 20 Dark counting rate: ~ 400 cps @ 25 °C, ~ 100 cps @ 15°C 20 Timing: ~ 160 ps many-photons regime, ~ 300 ps single photon regime Timing: ~ 160 ps many-photons regime, ~ 300 ps single photon regime Afterpulsing: ~ 10 -3 pulses for primary avalanche Afterpulsing: ~ 10 -3 pulses for primary avalanche Limitations: full recovery ~ 1,5  s full recovery ~ 1,5  s 5X5 Array performances (20  m) Dark Counts uniformity: ~ 10% Dark Counts uniformity: ~ 10% Cross-Talk: prompt ~ 10 -5 pulses for trigger Cross-Talk: prompt ~ 10 -5 pulses for trigger delayed ~ 10 -3 pulses for trigger delayed ~ 10 -3 pulses for trigger No distance dependence No distance dependence SiPM configuration: poor resolution SiPM configuration: poor resolution peak sensing (ADC) better then charge sensing (QDC) peak sensing (ADC) better then charge sensing (QDC) Conclusion 2007 - 1 st SiPM prototype ( ~ 5000 pixels), 1 st array for imaging 2007 - 1 st SiPM prototype ( ~ 5000 pixels), 1 st array for imaging