Impact of screening of latent defects at electrical test on the yield-reliability relation and application to burn-in elimination

This paper addresses the question of under what conditions burn-in can be eliminated. Based on data of more than 30 million sold devices, the effect of screening of latent defects at electrical test on product reliability has been investigated. The results are combined with the yield-reliability relation and an experimentally determined failure rate time evolution, yielding a model that allows determination of the sense or nonsense of burn-in or screens at electrical test quantitatively. The model predictions are in good agreement with experimental data. Furthermore, for typical operating conditions, high yielding batches show a better long term reliability than low yielding batches even if the latter have been subjected to burn-in. It is also shown that voltage stresses, distribution tests and IDDQ screens can be good alternatives to burn-in.

[1]  D. S. Peck New Concerns About Integrated Circuit Reliability , 1978, 16th International Reliability Physics Symposium.

[2]  Burton A. Unger Early life failures , 1988 .

[3]  N. F. Khory,et al.  A Rational Basis for Setting Burn-In Yield Criteria , 1984, ITC.

[4]  F. Kuper,et al.  Relation between yield and reliability of integrated circuits: experimental results and application to continuous early failure rate reduction programs , 1996, Proceedings of International Reliability Physics Symposium.

[5]  W. Smith,et al.  Does the burn-in of integrated circuits continue to be a meaningful course to pursue , 1988, 38th Electronics Components Conference 1988., Proceedings..

[6]  Wojciech Maly,et al.  Simulation of yield/cost learning curves with Y4 , 1997 .

[7]  D. L. Crook,et al.  Evolution of VLSI reliability engineering , 1990, 28th Annual Proceedings on Reliability Physics Symposium.

[8]  Timothy R. Henry,et al.  Burn-in elimination of a high volume microprocessor using I/sub DDQ/ , 1996, Proceedings International Test Conference 1996. Test and Design Validity.