Construction of the CMS Tracker End-Caps and an Impact Study on Defects

The CMS experiment at the LHC accelerator at the research center CERN close to Geneva will study proton proton collisions at up to now unprecedented centre of mass energies from the year 2008 on. To discover theoretically predicted elementary particles, CMS was equipped with the largest silicon tracker so far with a sensitive area of 198m2. Partitioned into more than 15.000 silicon strip modules, the construction and test of the tracker was a huge challenge for the involved institutes. The III. Physikalisches Institut B of the RWTH Aachen had a leading role in the construction and test of substructures, so called petals, for the end caps of the tracker. The petals were assembled in a clean room and underwent first basic tests to ensure the general operationability of each component. Failures detected during the assembly are described and improvements of the silicon strip modules are discussed. After the assembly the petals underwent a cold test for several days. For the first time all readout components of the petal were tested together at a temperature of −10C, similar to the final conditions in CMS. Efficient analysis methods were developed, that allow to find defects on the level of single silicon strips and hence obtain an overview over the quality of the components under thermal stress. The analysis of 294 assembled petals resulted in a single strip failure rate of less than 0.4%. Subsequently the petals were inserted in to the tracker end caps and shipped to CERN where further cold tests were done with both end caps. Presented here is the cold test of one of the two end caps which was constructed at CERN. The analysis of the failures detected is illustrated and the excellent quality of the tested end cap is confirmed. The measurements yield a failure rate of 0.4% comparable to the single petal test. The final chapter discusses the impact of defects on the track reconstruction. Defects were implemented into a MC simulation. Track reconstruction is done using two algorithms which are compared with each other. The analysis shows that even with a large number of defects as large as 5% a very good track reconstruction efficiency is obtained. Taking the known defect rate of May 2008 of 0.2% no effects on the track reconstruction efficiencies are observed.