Quality Assurance during production of the ATLAS18 ITk strip sensors

Le résumé fourni par la source

We describe the extensive quality assurance programme undertaken during the production of the ATLAS ITk strip sensors. Quality Assurance (QA) is the systematic process of preventing defects during production, rather than identifying them after the fact. QA is performed on test structures fabricated at the periphery of the same wafers used to produce the actual sensors, thus ensuring that these structures are a good representation of the sensor and process quality. The test structures consist of test chips, which include strips, diodes, MOS capacitors, and special structures for monitoring of parameters such as bias resistor values, interstrip resistance, coupling capacitance and flat band voltage. Additionally, miniature sensors with the same design as the main sensor, but with reduced area, allow measurement of the charge generated during irradiation with a beta source. Test structure semiconductor parameters are measured on both unirradiated samples and samples subjected to irradiation. The test chips are irradiated using gammas and reactor neutrons to the equivalent dose (660 kGy) and fluence ( 1.6 × 1 0 15 n e q / c m 2 ) expected in the high luminosity-LHC after 1.5 times its lifetime of 14 years. Miniature sensors are irradiated to the same neutron fluence as the test chips, or the equivalent proton fluence. Monitor diodes are irradiated alternatively to the maximum fluence to measure the degradation of the leakage current, and to one third this value to observe the full depletion voltage within the operational range of the measurement setups. Unirradiated test structures are measured on a probe station at room temperature. Irradiated test structures are wire-bonded to a PCB to provide connectivity, and measured at -21 °C and 4% relative humidity. After almost five years of sensor production, the QA process has shown reliable performance across multiple test sites. A few deviations from the acceptance criteria, such as low flat band voltage, high full depletion voltage, and variations in the magnitude and uniformity of the p-stop doping have been identified and timely reported back to the vendor. However, most of the evaluated parameters have remained within the defined specifications. We have tested ∼ 90% of the total sensor production to date. Of these, we have accepted 596 batches and rejected 6; giving a batch rejection rate of 1%. Of the six batches rejected, four are rejected due to QA issues related to low p-stop doping.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.