CMS pixel detectors

Results from First Tests with CMS Pixel Detectors

A silicon and a diamond sensor (DS-73), bump-bonded by Roland Horisberger to the PSI30 readout chip (Honeywell version) at PSI, have been studied. The setup of this measurement was developed at HEPHY, LabWindows/CVI has been used to analyse the data. The program was written by H. Steininger. You can take a look at a screenshot, where the analog data (red histogram), the digital information (blue histogram) and the control units are shown.

Pulse Injection

In order to determine the number of pixels connected, a voltage pulse U generated by an external pulser was injected into the backplane of the sensor. The response of the sensor was studied for various pulsheights and thresholds.

To estimate the charge Q=C*U injected into a single pixel, the capacity of a single pixel was calculated from the following equation
C = e_r * e₀ * A / (d*N) ,
with e₀ the dielectric constant, e_r = 11.7 (5.9) the dielectric constant of silicon (diamond), A=10.3125 mm² the detector area, d=300 um (600 um), the thickness of the silicon (diamond) detector and N = 660 the number of pixels. This gives a capacity of 5.5 fF for a silicon and 1.3 fF for a diamond pixel.

Results of the Threshold Scans

To estimate the sensitivities of the detectors, we varied the backplane voltage until a stable readout of a single pixel was possible. This was done for different thresholds. In the tables below we find the results of our threshold scans.

For a stable readout a 5 times higher voltage pulse had to be injected into the diamond sensor than into the silicon sensor, which is about 25 percent more than what we would expect from the different capacities. This we attribute to the combination of a low bump-bonding yield of only 3 pixels and the threshold spread in the readout chip. Therefore the probability, that the working pixels are the most sensitive ones is low. For the silicon sensor we saw a response of nearly all pixels to the backplane pulse and conclude that the the bump bonding yield was close to the the maximum.

The detailed description of the setup is shown in the following drawings and pictures, the full analysis can be found in the silicon and the diamond section.

Source Measurements

The detectors were exposed to a 37 MBq ⁹⁰Sr source. We could detect signals with the silicon pixel detector, but not with the diamond sensor. We attribute this to the low bump-bonding yield (resulting in a low probability to hit one of these pixels) and the for diamond sensor relatively high threshold.

Aug 2 1999, Thomas Bergauer, Rudolf Wedenig