Overestimation of NBTI lifetimes measured on hot chuck for PMOS with antennas for PID assessment

Reliability data of three different stress methods are compared for a PID characterization of pMOS with various antenna sizes. In focus is the conventional NBTI utilizing a hot chuck for which anomalous data were reported previously with smaller NBTI degradation when plasma damage is present. However, it is shown here that the data of an NBTI stress with a local heater is more trustworthy. Furthermore, this NBTI degradation data can be correlated successfully to data of a fast WLR PID stress at room temperature. Trapping characteristics during NBTI recovery at high stress temperature for the conventional NBTI stress are found to be the root cause for the “reverse antenna effect”.