Diagnosing IC failures in a fast environment

Abstract

In today's world of fast product introductions, short product life cycles, and complex ICs, manufacturers need comprehensive, accurate, and fast diagnosis of development failures and reliability hazards before launching a product. As a result, the failure analyst's role has changed from reactive to proactive. Waiting until failures come back from the field to assess the reliability of a product is no longer efficient. By that time it's too late. Devising tests that anticipate all failure modes and accelerating those failures within a reasonable length of time is also impractical if not impossible. There are too many variables and too many unknowns in the equation to use this shotgun approach. It results in too many reliability tests that produce no failures, and failure rates that too often are calculated on the basis of potentially irrelevant tests. To adequately evaluate the reliability of a component, the proactive failure analyst must first closely inspect it for both physical and electrical hazards. Combining this data with field data from products using the same technology, the analyst can run realistic tests that stress the part to failure, thus deriving an appropriate failure rate. The proactive failure analyst's role is to detect failures before they occur and assist the circuit designer in eliminating their cause before the product ships. This means that the analyst must conduct evaluation testing with an eye toward the physics of component failure. The result is improved manufacturing yield and a more robust, more reliable component in the field.