1 / 24

Next generation 3D digital SiPM for precise timing resolution

Next generation 3D digital SiPM for precise timing resolution. Jean-François Pratte F Nolet, W . Lemaire, F. Dubois, N. Roy, S . G. Carrier, A. Samson , G. St-Hilaire, S. A. Charlebois, R. Fontaine.

michelm
Download Presentation

Next generation 3D digital SiPM for precise timing resolution

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Next generation 3D digital SiPMfor precise timing resolution Jean-François PratteF Nolet, W. Lemaire, F. Dubois, N. Roy, S. G. Carrier, A. Samson, G. St-Hilaire, S. A. Charlebois, R. Fontaine Interdisciplinary Institute for Technological Innovation (3IT), Université de Sherbrooke 14th Frontier Detectors for Frontier Physics 2018

  2. Introduction Time of flight for PET • Δ time = Δ distance Time-of-Flight Stop Time-to-digitalconverter + Patient • PET Scanner - Start • 10 ps coincidence timing resolution (CTR) => 1.5 mm of precision

  3. SNR gain as a function of the Coincidence Timing Resolution • In our case, a 10 ps CTR would represent a CNR gain of >10 in the image • Could lead to real time imaging in PET • Many challenges: • Scintillator with prompt photons • Photodetectors and electronics with 10 ps SPTR (Single Photon Timing Resolution) P.Lecoq, 2017, IEEE TRPMS

  4. Single Photon Avalanche Diode (SPAD) Operation Cycle 3: Recharge A C 1: Trigger 2 3 2: Quenching 1 B Excellent single photon timing resolution Sensitivity – single photon counting Silicon  mass production  low cost

  5. Analog Versus Digital SiPM… a Paradigm Shift… Bottom line, with an analog SiPM we sum binary detectors (an array of SPAD) to get a linear response… Then, use a current/transimpedance amplifier + shaper + ADC to digitize the data… again! • How to take advantage of the SPAD’s digital nature? • Can power consumption be saved using a digital approach compared to the classic analog readout dealing with large input capacitance of the serial / parallel connections of SiPM? jean-francois.pratte@usherbrooke.ca

  6. Benefits of Digital SiPM for TOF-PET 2D Digital SiPM • Advantages • 1 TDC per SPAD • Uniform SPTR per pixel • SPAD to SPAD skew correction • Cons of 2D implementation • Low fill factor • Sameprocess for SPAD and CMOS • No room for digital signal processing Digital SiPM

  7. Sherbrooke’s 1st Generation 3D digital SiPM • CMOS readout: Teledyne-DALSA 0.8 µm HV • SPAD: Modified CMOS process

  8. NextGeneration 3D Digital SiPMwith CMOS 65 nm readout 3D Integration • High fill factor • Heterogeneous technologies integration: • SPAD array: Teledyne-Dalsa custom process • TSMC CMOS 65 nm - 256 SPAD readout ASIC

  9. 3D Digital SiPM readoutarchitecture 256 pixels • Informations per packet: • Pixel address • Timing of the first photon detected • Photon count over period of time • Array trigger modes: • Time driven • Camera mode: timing of the first photon + photon count at a fixed time rate • Event trigger (i-e: TOF-PET) • X columns must trigger to detect an event • Timing of the first photon + photon count for a programmable time window SPAD SPAD SPAD SPAD QC QC QC QC TDC TDC TDC TDC Pixel Pixel Pixel Pixel C) Dark count filter Array Readout A) Calibration and correction B) Timestamp sorting D) Multi-photon estimation

  10. Pixel overview 65 µm

  11. On-chip digital signal processing– Raw Data 256 pixels 1st Mode of data transmission Raw data information • Raw data (address, timestamp, count) • Corrected raw data • TDC LSB correction (look-up table) • Skew correction (look-up table) • Measured bandwidth 14 kcpsfor 1.1 × 1.1 mm2 SPAD QC TDC Channel Out 1 2) Calibration and correction 3) Timestamp sorting 4) Dark count filter 5) Multi-photon estimation Array Readout Global Counter 1) Raw data Out 2

  12. On-chip digital signal processing– ProcessedData 256 pixels 2nd Mode of data transmission • Multi-photon time estimator based on BLUE (Best linear unbiased estimator) • Sort the 32 first timestamp • Filter dark count • Output is: • Timing information of 32 pixels combined • Energy (counts) sum of 256 pixels • Measured bandwidth0.5 Mcps for 1.1 × 1.1 mm2 SPAD QC TDC Channel 2) Calibration and correction 3) Timestamp sorting 4) Dark count filter 5) Multi-photon estimation Array Readout Global Counter 1) Raw data Out 5

  13. What are the measured contribution to the SPTR ? Timing jitter contribution within the array • Timing jitter • SPAD • QC • TDC • Array level contribution • Noise between pixel • TDC LSB non-uniformities • Pixel-to-pixel skew • N/A since the digital SiPM is not yet integrated in 3D • < 3 ps rms with a pulse generator at its input • 8 ps rms for a TDC LSB of 15 ps

  14. Timing Skew and Jitter – ArraylevelMeasurements • Skew contribution: clock tree, trigger tree, pixel mismatch • Skew max of 500 ps within 1.1 1.1 mm2 • Total Jitter is 87 ps rms

  15. Timing jitter – Arraylevel • Since each pixel address is known, it can be characterize and corrected • Skew corrected (1 ps max skew) • Total Jitter is 18 ps rms

  16. Impact of the number of active TDC on the jitter What’s in the ASIC • Triple well isolation • Separated power supplies for TDC – QC – digital core • Analog buffer in each pixel • The jitter worsen from 2 ps to 5 ps • For 256 TDC • The jitter worsen of ~1 ps • For 1 to 64 TDC (less common noise)

  17. BLUE - preliminaryresultswith on-chip readout • Simulation with Geant4 for LYSO (1×1×3 mm3) scintillatorstatistic • Loadparameter in Cadence to see the impact in Simulation • Load the parameter in the ASIC and see the impact of the TDC jitter and the post-processing • Difference = Higher TDC jitter Simulation Measurement

  18. Conclusion • We developed Digital SiPM readout in CMOS 65 nm for Time-of-Flight Measurements with one QC and one TDC per pixel • With the implemented correction circuit, the timing jitter is 18 ps rms, close to our objective to 10 ps array wide. • We are currently correcting the analog buffer and ring oscillator • Each step of the on-chip digital processing are fully functional • Next revision focus • 4 : 1 TDC • Non-uniformities correction • Power optimization • Calibration system

  19. Conclusion • 3D digital SiPM is a versatile technology. • A 3D digital SiPM iscurrentlyunderdevelopment for LAr and LXeexperiments • We are member of the nEXO collaboration for double beta decaysneutrinoless • Nextgenerationhundred tons LAr

  20. Acknowledgements • Thankyou for your attention

  21. Back-up slides

  22. Digital SiPM Overview SPAD SPAD SPAD SPAD QC QC QC QC TDC TDC TDC TDC 256 Pixels Channel Channel Channel Channel Array Readout A) Calibration and correction C) Dark count filter D) Multi-photon estimation B) Timestamp sorting Out 1 Out 2 Post-processing Out 3 Out 4 Out 5

  23. Dark count filter Multiple windows filter Array Readout 2) Calibration and correction 3) Timestamp sorting 4) Dark count filter 5) Multi-photon estimation Global Counter 1) Raw data PLL (x2) Dark noise Event

  24. 3D digital SiPM Analog Monitor for Physicist Transimpedance+  of 484 channels

More Related