Development and Qualification of Copper Wire Bond Process for Automotive Applications

摘要

This paper provides a detailed study on the challenges on development of a robust Cu wire bond process on thin aluminum (Al) bond pad (0.8μm thickness) to meet Automotive Electronics Council Grade 1 requirements (AECG1). The test vehicle used in the study was a 0.18μm technology device assembled in a 100 leads, 14mm × 14mm Quad-Flat Package (QFP). The device has a tungsten sea of vias design beneath the Al pad metallization, which renders the device to be highly susceptible to pad cratering on Ti/TiN barrier metal layer. This is especially true for Cu wire bonding, where the hardness of Cu wire and requirement to achieve good intermetallic coverage (IMC) formation between Cu and Al is extremely critical. Hence a robust wire bond process optimization is very important. Wire bond process optimization is performed through several rounds of design of experiments (DOE) and response surface methodology (RSM) to establish the bonding parameter window by optimizing various wire bond responses such as bonded ball size, bonded ball thickness, wire pull strength, ball shear strength, intermetallic coverage, cratering test, as well as post-thermal aging wire pull testing. Samples were fully assembled into singulated units from 3 different wafer lots, with total of 5 assembly lots to cover the wire bond process window range at Low, Nominal and High setting. The samples were then subjected to reliability stressing as required by AECG1 qualification, such as High Temperature Storage Life (HTSL), MSL3/260°C pre-conditioning + biased Highly Accelerated Stress Test (HAST) and MSL3/260°C pre-conditioning + Temperature Cycling. Results showed that with a well optimized wire bond recipe, Cu wire bond process was able to meet at least twice the AECG1 requirements for the various reliability stress conditions. A portability study on multiple wire bonders was also carried out to ensure the robustness of the process towards high volume manufacturing.