The HERA-B Vertex Detector System

The Verdex Detector System [VDS] was built in and around a vacuum vessel which held 7 super-layers of silicon vertex detector modules.

Each super-layer had 4 quadrants with 2 double-sided silicon strip detectors each.

The MPI Munich and MPI-K Heidelberg groups shared the responsibility for the construction of VDS.

Each detector was mounted on a module, additionally carrying the front end readout electronics. These detector modules were designed and produced by the Munich group, while the Heidelberg group was in charge of the VDS infrastructure, including the readout system and the vacuum vessel.

The VDS worked well within design specifictions during the duration of the HERA-B experiment. Please check the publication section.

The HERA-B Silicon Detector Modules

The design of the modules was started in summer 1994 and the main installation of the modules happened in 1999. Several prototypes were tested in the lab and in the experiment itself. The first prototypes were installed in spring 1996. They were based on existing preamplifier chips and simplified mechanices.

The final modules had especially developed preamplifier chips and a special mechanical design that facilitated different cooling paths for detectors and electronics. Two modules were mounted back to back.

The modules were constructed out of silicon detectors, electronics hybrids, flexible pitch-adapters called flex-jumpers and carbon fiber plates mounted on cooling blocks.

Each silicon strip detector had 1280 channels on the front and 1056 channels on the back-side. Thus about 5000 bond wires had to be attached to complete a module. The flex-jumpers were long enough to mount the electronics far enough from the beam to eliminate radiation damage. The design allowed for up to 10MRad at the edge of the detector close to the beam. The electronics would have been exposed to 100kRad in the same period. All modules were removed in 2004 when the HERA-B experiment was disassembled. None of the modules ever saw the design dosage.

The strips of the silicon detector were rotated by 2.5 degrees. The back to back mounting of two detectors per super-layer thus provided four stereo views.

Carbon fiber was used as a material in order to reduce the material in the VDS. Material reduction is essential for all forward systems.

The first prototypes

The first prototypes installed in 1996 had a much simpler mechanical design based on a single ceramic carrier. They helped us gain the experience to design the final product.

The pitch-adapters were also made of ceramics. The strips of the silicon detectors were not yet rotated.

It takes many a part to make a module

Many production steps are needed to make a module. The silicon detectors were designed at the semiconductor lab HLL operated by the MPI of Physics and the MPI of Extraterrestrial Physics. The detectors were partially produced by the HLL and partially by SINTEF in Oslo. The ceramic hybrids for the electronics were produced at the MPI. The pre-amplifiers were designed by the ASIC lab of the MPI-K and produced commercially. The flex-jumpers were produced by Dyconnex. The bonding was done at the Rutherford Appleton Laboratory [RAL] and during crunches also at CERN.

All gluing and assembly work was done at the MPI in Munich. The gluing of the silicon detectors proved to be critical as the wrong glue can increase the leakage current significantly. After long tests a silicone paste was selected. The silicone provides a relatively soft connection between the carbon fiber and the silicon detector minimizing elastic stress due to the different thermal extension coefficients of the two materials.

Under normal circumstances it took about three months to produce a module and we finished about one module per week. During production the modules were shipped frequently. Some of them got their own frequent traveler cards from Lufthansa.

Quality Control

The strips of the silicon detectors were tested for shorts to ground and against each other. This was done by the manufacturers using probe-stations. A maximum of 10 damaged strips per detector side was accepted.

The electronics hybrids were tested at the MPI in Munich before they were integrated in modules. Only two dead electronics channels per hybrid [detector side] were accepted.

The flex-jumpers were tested by the manufacturer. No disconnected traces were allowed.

Between different production steps IV-curves were taken. This assured that the detector performance was not changed by the integration into a module.

After the completion of a module the depletion voltage was measured and all strips were checked for their performance. This was done using the specially designed laser test stand . Complete lists with dead and noisy channels were thus provided for data analysis in the experiment.

Last modified: Mon Sep 27 19:25:13 MET 2004