Title :
A Statistical Analysis of the Number of Failing Chips Distribution
Author :
O´Donoghue, Geoff ; Gómez-Uribe, Carlos A.
Author_Institution :
MediaTek Wireless, Inc., Wilmington, MA
Abstract :
We find that the number of failing chips per wafer has a probability distribution with a consistent structure across multiple technologies and product types. We further show that a general extreme value provides an accurate and efficient means of describing this distribution. We also show that chips on a wafer tend to fail together, particularly if they are physically close to one another; this behavior results in clusters of failed chips on a wafer. A method for the direct measurement of clusters of failing chips on a wafer is described, along with a model to relate this measurements to the distribution of the number of failing chips. The methods described here are potentially applicable for identifying wafer outliers or for monitoring of the manufacturing process.
Keywords :
integrated circuit manufacture; integrated circuit reliability; statistical distributions; failing chips distribution; manufacturing process monitoring; probability distribution; statistical analysis; wafer outliers; Condition monitoring; Helium; Manufacturing processes; Probability distribution; Semiconductor device manufacture; Semiconductor device measurement; Semiconductor device modeling; Shape; Statistical analysis; Testing; Clustering; faults; number of failing chips (NFC); yield;
Journal_Title :
Semiconductor Manufacturing, IEEE Transactions on
DOI :
10.1109/TSM.2008.2001208