Process Optimization of a Selective Microstructured Nickel/Gold Metallization aimed at Topology Adaptation for the Hybridization of ASICs and Silicon Pixel Sensors

Hybrid silicon pixel detectors require a reliable interconnect technology to electrically and mechanically connect pixel sensors and readout electronics. As an alternative to conventional solder-based hybridisation, this work investigates a novel in-house hybridisation approach based on Anisotropic Conductive Adhesives (ACAs), which requires a specific protruding pixel pad topology. This topology is realised using a selective Elektroless Nickel Immersion Gold (ENIG) plating process on aluminium pixel pads. The focus of this thesis is the systematic optimisation of the ENIG process with respect to bump height, uniformity, and reproducibility. A detailed parameter study is conducted, analysing the influence of pretreatment, in particular the number of zincations, as well as nickel deposition temperature, pH value, stabiliser concentration and deposition time. The results show that the pretreatment is a key factor for achieving homogeneous nickel growth. Multiple zincations significantly improve surface uniformity and reduce roughness of the ENIG bumps. The nickel deposition rate is found to be parameter dependent in a range of 3-26 μm/h and can be precisely tuned by adjusting temperature and pH. Inappropriate stabiliser concentrations are shown to cause either parasitic nickel deposition on the passivation layer or premature termination of growth at pad edges. Based on the identified optimal process window, uniform ENIG bumps with well-controlled height are reproducibly achieved for samples without prior defects. Flip-chip assemblies using ACA and daisy-chain test structures demonstrate a high connection yield of up to 99%, confirming the suitability of the optimised ENIG process for fine-pitch hybridisation.The presented results show the potential of ENIG plating in combination with ACA as a flexible in-house solution for the hybridisation of pixel detectors, particularly for small-volume prototyping in research environments.