Using Volume Cell-aware Diagnosis Results to Improve Physical Failure Analysis Efficiency

Abstract

Statistical analysis based on layout-aware scan diagnosis has been successfully used for identifying defect root causes and reducing physical failure analysis (PFA) efforts, especially for interconnect defects. With increasing complexity and density of designs manufactured by FinFET technologies, more and more cell internal defects are observed. For such defects, the root cause deconvolution (RCD) and PFA based on layout-aware diagnosis learning become less efficient because diagnosis reports can only call out cell instances, but can’t pinpoint the defect location within the suspected cell. Cell-aware diagnosis (CAD) uses analog simulation results to accurately locate defects inside standard cells. The cell-aware RCD (RCAD) provides a comprehensive defect pareto for both cell-internal defects and interconnect defects. Both techniques can be very beneficial for PFA. In this work, we present a case study which combines these techniques to successfully identify a systematic cell internal issue caused by a sensitive layout pattern with dramatically improved PFA efficiency for recent silicon data manufactured by an advanced FinFET technology.