R. E. Mulder, Freescale Semiconductor, Austin, TX

Summary: The establishment of nano-probe technology as a failure analysis tool has given product analysts the ability to accesses all the nodes of any transistor on a chip allowing for the localization and complete electrical characterization of a defective transistor. This ability has resulted in a significant increase in number of individual soft transistors defects found as the cause of the device failure. These soft transistors defects are usually seen as Vt shifts, suppressed drive currents, or low level leakages. Usually, the cause of these soft defects is not identifiable by means of physical analysis such as the Scanning Election Microscope(SEM), Transmission Electron Microscope(TEM), or Atomic Force Microscopy(AFM)[1]. Therefore, other methods are being sought out to identify the root cause of transistor failures. Transistor modeling of possible defects using TCAD and comparing results to passing and failing transistor characterization is promising [2]; but accurate transistor modeling isn’t easy and is resource intensive. Another alternative is to figure out ways to identify more information out of the electrical characterization data that will indicate the possible root cause of the failure. This paper is a case study that starts the process of exploring alternative methods of electrical characterization and analysis of the data that will help identify the root cause failure mechanism for soft transistor failures.