Timing in Thick Silicon Detectors Andrej Studen University

Timing in Thick Silicon Detectors Andrej Studen, University of Michigan, CIMA collaboration

Outline Motivation Timing in pad detectors Two intuitive solutions Comparison to measured data Where to go from here

Motivation Thick silicon detectors improve efficiency for gamma-ray detection. In a coincident setup (PET, Compton camera) good timing resolution is required. Experimental data not promising [1]. Could it be compensated by different readout strategy and bias conditions? [1] N. Clinthorne et al. Timing in Silicon Pad Detectors for Compton Cameras and High Resolution PET; IEEE NSS/MIC, Portoriko, 2006

Model application Silicon pad sensors used in Compton & silicon PET experiments at Uof. M: n n n p+-n-n+, 256/512 pads Pad size 1. 4 x 1. 4 mm 2, Thickness: 1 mm, FDV: 150 V (!), ASIC: VATAGP 3: w Charge sensitive pre-amplifier w CR-RC shaper with 200 ns shaping time. w Leading edge discriminator.

Signal formation in a pad detector Readout electrode Signal shape depends on: Electric field Ramo field Interaction depth P-side holes Recoil electron ~100 um (E electrons gamma) N-side + Charge q moving in electric field gamma-ray induces current pulse on readout electrode: Compton scattering or photo-absorption

Electric field Thickness (1 mm) « lateral dimension (12/24 mm by 48 mm). P-n junction. Large field region. harges are fast. Low field region. Charges move slowly.

Ramo Field Pad size ~ depth. Asymmetry. Large Ramo field. Large contribution to current pulse. Low field region. Charges contribute less.

Interaction depth Two regions: n Near region: w Large E field, w Large Ramo, w Fast rise-time. n Far region: w Small E field, w Small Ramo, w Slow rise-time. Very sensitive because of short electron path. Asymmetry of both fields works against us.

Example: single e-h pair at pad edge, 1. 4 VFD Trigger time shift Far region f. Z=0. 9 Leading edge trigger Detector Near region f. Z=0. 1 Preamp, CR -RC; t=200 ns

Solution 1: Adding adjacent pads Reducing Ramo asymmetry. Noise of 9 pads added – jitter increased 3 x

Solution 2: Increasing bias Much shorter times w/ higher bias Often unpractical

Simulation overview GEANT 4 used to generate “true” paths of recoil electrons 661 ke. V photons; 137 -Cs (also measured) Voltages from 200 V -> 400 V Both single and summed pads

Results overview Threshold: n 15 ke. V (experiment). Time-walk: n n n Dominates below ~ 100 ke. V: Could be compensated by appropriate readout strategy. Three levels assumed for illustrative purposes.

Comparison to measurements Measured in Compton mode (PMT start, silicon stop; PMT timing resolution ~ 10 ns) Sharp edge Blunt edge Spurious tail

Comparison, U=400 V Simulation marginally better, measurement data more symmetric. Spurious tail gone.

Solution simulation RAMO 9 pads 200 V BIAS 400 V 1 pad

Latest greatest Do both!

Conclusions Shape of Ramo field has a significant influence on timing in thick silicon detectors. Solutions: n n Multi-pad readout (noise!), Different detector geometries (strips? ) Different trigger strategies. Operate at higher bias voltages.

Backup slides subtitle

Illustration of depth-related trigger time variation 15 % trigger, 1. 4 VFD Single pad 9 pad sum 70 60 50 40 ns ns 30 ns 20 ns

Illustration (cont’d) 15% threshold, 1. 4 VFD Single Pad 9 pad sum