Abstract
Several hundred units were subjected to autoclave stress as part of the qualification of a new fast static RAM. Many units failed after autoclave stress, and these parts recovered after conventional depotting using nitric acid and a hot plate. Based on the recovery of the units, the failures were determined to be fuse-related because the nitric acid cleared the fuse cavities during depotting. Chemical analysis after thermally extracting the die from the package revealed an antimony-rich material in failing fuse cavities. Source of the antimony was linked to antimony trioxide added to the plastic package as a fire retardant. However, it was unclear whether the antimony-rich material caused the failure or if it was an artifact of thermal depotting. A new approach that did not thermally or chemically alter the fuse cavities was employed to identify the failing fuses. This approach used a combination of back-side grinding, dimpling, and back-side microprobing. The antimony-rich material found in the fuse cavity was confirmed using SEM and TEM-based EDS analysis, and it is believed to be a major contributing factor to fuse failures. However, it is unclear whether the short was caused by the antimony-rich material or by a reaction between that material and residual aluminum (oxide) left in the fuse cavity after the laser blows.
DEVELOPMENT OF THE FABRICATION METHOD FOR STATIC RAM RADIATION HARDENED SILICON-ON-SAPPHIRE MICROCIRCUITS
