Charge collection studies with irradiated CMOS detectors

15x6 array of 50x250 um2 pixels Contacts: single pixel all other pixels Structure B IR laser beam Two sets: unthinned (700 um), no back plane, substrate bias over implant on top thinned to 200 um, back plane processed, bias through BP

Passive test structure B, substrate resistitivity 300 kOhm cm

E-TCT, not thinned samples Bonded: outmost ring (substrate) – to GND all n-wells except single single n-well  HV connected to n-wells Charge profiles at different bias Single n-well read out (others at HV) Profile width vs. bias

E-TCT after irradiation not thinned samples

Neff vs. Phi Zoom to low fluences N_eff smaller than before irradiation not measured  probably below 1e13, should measure there to get good estimate of acceptor removal constant linear fit good enough, g_c larger than “usual” value g_c = 0.02 cm-1 similar as previous LFoundry

Measure also samples thinned to 200 um samples full depletion clearly visible on thinned sample

90Sr setup HV-CMOS: small signals, large noise  S/N very low detector ORTEC 142 charge sensitive amplifier + custom made shaper (25 ns) + digital oscilloscope 90Sr PMT ampl. cooling jig DUT HV-CMOS: small signals, large noise  S/N very low clean sample of events needed (no hits missing DUT) require a large detector (trigger rate), good collimation, small scintillator Measurement: Calibration with a 300 µm thick Si pad detector 1) Record N ( = 2500) waveforms 2) Average over all waveforms and determine time of the signal peak 3) Sample waveforms at the peak 4) Fill spectrum

depleted depth from E-TCT ~ 280 um  expected MPV ~ 21000 el Structure B is small and we can’t collimate so that we measure only events with tracks passing through the detector but before irradiation and at low fluences signal and noise peak well separated and Landau could be fit Before irradiation: MPV = 13300 el 260 V Thinned Thinned: Depleted depth ~ 190 um  expected charge 13500 el Not thinned: depleted depth from E-TCT ~ 280 um  expected MPV ~ 21000 el Consistent!

After irradiation, Ф = 5e13 n/cm2 MPV = 13100 el MPV = 9000 el Thinned to ~ 200 um, substrate bias via back plane depleted depth from E-TCT ~ 180 um (fully depleted)  expected MPV: ~ 13500 el  measured MPV ~ 13000 el Not thinned, substrate bias from top: depleted depth from E-TCT: ~ 260 um  expected MPV ~ 19000 el  measured MPV ~ 9000 el  Large difference between samples with and without back plane after irradiation! Diffusion neglected (from measurements with AMS structure we expect ~ 1000 el before irradiation (~ 0 el after irradiation))

Much smaller charge than expected from E-TCT in samples with top substrate bias Samples with Back Plane contact as expected

Explanation: difference due to different weighting field in back or top bias after irradiation No back plane, substrate biased via implant on top Back plane (and thinned), substrate biased via back plane Collecting electrode: Uw = 1 Collecting electrode: Uw = 1 Depleted region Depleted region e d D d h Undepleted bulk Substrate bias: Uw = 0 Before irradiation: Undepleted substrate before irradiation low resistivity (Ohmic) weighting potential Uw = 0 carriers drift across whole weighting field: all charge collected After irradiation: substrate resistivity (Ohmic) high weighting potential 0 at the back plane implant  if fully depleted D = d full charge collection (except trapping)  if not fully depleted carriers don’t cross whole weighting field  charge collection reduced by a factor d/D  depending on geometry and device thickness this factor can be much better than in the case of top bias Before irradiation: Undepleted substrate before irradiation low resistivity (Ohmic) weighting potential Uw = 0 carriers drift across whole weighting field: all charge collected After irradiation: substrate resistivity (Ohmic) high, weighting potential 0 at the bias implant on top  Carriers trapped in low field at the end of depleted depth, before they reach the substrate bias electrode  carriers don’t drift across all whole weighting field  partial charge collection