Relation between yield and reliability of integrated circuits and application to failure rate assessment and reduction in the one digit FIT and PPM reliability era

IDDX-based test methods

“…Moreover, BI is getting prohibitively expensive. Research has shown that chips having higher I DDQ also have poor long-term reliability [Van Der Pol et al 1996]. The potential of I DDX test to eliminate [Henry and Soo 1996;Kawahara et al 1996;Wallquist 1995b], or reduce BI has been investigated [Mallarapu and Hoffman 1995;Righter et al 1998;Sabade and Walker 2002b].…”

On the Relationship of Semiconductor Yield and Reliability

“…Previously, there have been some attempts to extend yield models to estimate reliability. At first, a simple time-independent Poisson reliability was obtained from the Poisson yield model, assuming directly that the number of nonfatal defects is proportional to the number of fatal defects following a Poisson distribution [6], [13], [18], [19]. Later, the Poisson reliability was extended to the negative binomial reliability to consider defect clustering [2].…”

Modeling manufacturing yield and reliability

“…Another report for the strong relationship between yield and reliability is presented by Van der Pol et al [23]. Their research shows that a strong and measurable relationship exists between the number of failures in the field as well as in life tests and the yield due to the adoption of the wafer level reliability (WLR), and the use of reliability related design rules [23]. Thus, the root causes of reliability failures are the same as those of yield failures, and the manufacturing yield depends upon the number of defects introduced during the manufacturing process, which in turn determines reliability.…”

“…Prendergast [16] points out a linear relationship between yield and reliability and suggests that this relationship can be effectively used to screen unreliable products. Another report for the strong relationship between yield and reliability is presented by Van der Pol et al [23]. Their research shows that a strong and measurable relationship exists between the number of failures in the field as well as in life tests and the yield due to the adoption of the wafer level reliability (WLR), and the use of reliability related design rules [23].…”

“…In order to use the model of (1), it is necessary to calculate the reliability critical area. Kuper et al [13] and Van der Pol et al [23] use another model for the yieldreliability relation and present the experimental data to show the correlation. Their relation model is given by (2) where is a parameter for clustering effects and edge exclusions and is the ratio between the density of reliability defects and the density of yield defects .…”