Status of Indium Bumped FEI 4 B Richard

Status of Indium Bumped FE-I 4 B Richard Bates, Craig Buttar, Kate Doonan, John Lipp, Thomas Mc. Mullen 28/03/2014

Introduction RX Delay • 300μm CPII sensor Indium bump bonded to FE-I 4 B ROC • Set up on probe station in GLADD to readout using USBPix Analog • RX Delay and Analog and Digital scans done to test module was working • Tuning performed on the module to ascertain number of disconnected/merged bumps and dead pixels Digital

Disconnected pixels • (criteria: Sigma > 0 e and < 200 e) • This is the case for 6 out of 26880 pixels

Merged Bumps • Merged pixels • (criteria: Crosstalk occurs at 100 V) • This is the case for 106 out of 26880 pixels

Dead Pixels • As tuning was not ideal, this is an approximate value by counting pixels with sigma value less than 1 e from Threshold Scan @ 100 V from long tuning attempt (see later) • 21 pixels dead

Bump Yield • DISCONNECTED: 6 • MERGED: 106 • DEAD: 21 • TOTAL: 133 out of 26880 • BUMP-YIELD: 99. 505%

Tuning • The IBL tuning scheme did not work fully as it uses FAST tunes and all parameter space was not scanned • 3000 e threshold with 9 To. T @ 16 ke • THRESHOLD: Slightly high compared to target. White pixels here are not dead pixels, but pixels which are not ‘reached’ using a fast tuning.

Tuning • Compared to SC-3072 -4 -8 (250μm x 50μm) • THRESHOLD: 2975 e

Tuning • Sigma on threshold with 0 V bias shows very few pixels with values around 120 e (disconnected) • Next slide has log Y plot to show pixels with low values of sigma

Tuning • Log Y Scale on sigma plot

Tuning • Compared to SC-3072 -4 -8 (250μm x 50μm) at 100 V • SIGMA: 119. 6 e

Tuning • TOT: This looks good!

Tuning • Compared to SC-3072 -4 -8 (250μm x 50μm) • TOT: 8. 99

Tuning • LONG tuning at 100 V looked to be problematic: • Threshold very high • although note the pixels which were white in fast tune are occupied now • Sigma very spread out • Closer look at individual SCurves of pixels showed very poor fit giving both: • HIGH THRESHOLD! • LARGE SIGMA! HIGH THRESHOLD! SCurve is evident – fit has failed

Tuning • LONG tuning at 100 V looked to be problematic: • Threshold very high • Sigma very spread out • Closer look at individual SCurves of pixels showed very poor fit giving both: • HIGH THRESHOLD! • LARGE SIGMA! Poor fit – sigma actually much smaller!

Tuning • I am contacting Malte Beckhaus at CERN to see if there is a method of improving the fit for the SCurves • With this issue, the threshold value and sigma value for each pixel is wrong in this case