Browse Prior Art Database

Reliability Monitor Method for Early vs Late Fails

IP.com Disclosure Number: IPCOM000234580D
Publication Date: 2014-Jan-20
Document File: 4 page(s) / 88K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a Method to Predict Reliability Throughout the Life of a Product

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 4

Reliability Monitor Method for Early vs Late Fails

Current methods used to create reliability models do not account for variation in individual reliability mechanisms from beginning of life (BOL) to end of life (EOL). EOL fail rate is set to worse case reliability for each mechanism and there is no way to update reliability models after product qualification.

This novel method creates a fail rate for each reliability mechanism and mathematically combines all the mechanisms to provide a more accurate reliability estimate at any point in the the useful life of the product.

Figure 1 shows the bath tub curve currently is used to predict product reliability. BOL fail rate is dominated by manufacturing defects. EOL fail rate is impacted by reliability mechanisms such as electromigration (EM) and time dependent dielectric breakdown (TDDB)

Figure 1: Bathtub Curve Reliablity (BOL to EOL)

Figure 2 shows targets for EM and TDDB throughout the useful product lifetime. In this example (for illustration purpose only), EM performance is better than target tand TDDB is worse than target. Composite reliability at any point in the useful product lifetime will be the composite of all reliability mechanisms.

Figure 2: Expected Reliability Mechanism Fail Rate Compared to Target Reliability

1


Page 02 of 4

Figure 3 shows an example of EM and TDDB as a function of process sigma. EM, TDDB and some other reliability mechanisms vary as a function Performance Process Window (how fast or slow an individual part is). Reliability mechanisms that vary as a function of Performance Process Window include:


• Bias Temperature Instability (BTI), positive and negative, which varies as a

function of the Front End of Line (FEOL) Process Window; more fails at the slow end of distribution


• Hot Carrier Injection (Hot-e), which varies as a function...