Probabilistic Sensitivity Analysis (PSA) introduces randomness or variability into input parameters to reflect the uncertainty associated with those parameters. The primary purpose of PSA is to assess the overall uncertainty in model outputs by considering the joint distribution of input parameters. In PSA, random samples are drawn from probability distributions assigned to input parameters. Monte Carlo simulation is often used to generate many scenarios by sampling from these distributions. PSA Provides a distribution of model outputs, allowing for the estimation of the probability of different outcomes.
This presentation will show how PSA can be used to determine the impact of operational, environmental, and geometric factors, as a group, and individually to ensure the selection of parts for safe and reliable operation of systems. In addition, the presentation will show the influence of each input’s variability on the variability of the output, which is a measure of the reliability of the part. The information about interactions between inputs and statistical variability of the results is covered. The concentration of this presentation will be on the geometric parameters of semiconductor parts accounting for the inherent manufacturing and process variations. The presentation will introduce the concept of process family and related terminologies such as node size, feature size, and process node and show how the information can be used for reliable product development and deployment.
There are inconsistencies in terms of how dimensional parameters are presented based on process family. Also, the die-level dimensions of a part for the same node size vary from manufacturer to manufacturer and for the same manufacturer. Therefore, the effects of uncertainty in dimensions and the inconsistency in terminologies need to be addressed, and the effect of uncertainties of input parameters (dimensional parameters specifically for this presentation) on the “life estimation” methods of the part will be presented.