Understanding defects, diagnosis and variation

Lecture

May 22 10:30
Nidaros Pilgrim Centre

This tutorial gives a background on the nature of test in submicron processes, where a significant effort is devoted to measuring and understanding process variability, and care must be taken to choose where defect density leaves off and device variability begins. Three key technologies involved in this are diagnosis identifying failures and mapping them to physical causes, statistical design practices, which are used to tolerate expected amounts of variability, and test, which is used to eliminate variability that is outside of expected amounts.

Prerequisites & suggested preliminary readings

Basic VLSI design, circuit design, test/ATPG

Learning outcomes

Attendees will learn the basics of sub-65nm process variability, DFM, fault diagnosis, and statistical design. Emphasis is placed on practical application over theoretical background, although the latter is provided where needed.

Syllabus

  1. Introduction and background
    • Class goal
    • What is DFX? Interaction of manufacturability, yield, variability, reliability and test
    • DFM background
    • Photolithography basics
    • Resolution enhancement technology
    • Physical variability
    • Defects and fault models
    • Manufacturability versus yield
    • Economics of yield
    • Classes of yield (systematic, defect-related, parametric, design-related)
    • Yield models and metrics
  2. Diagnosis
    • Correlation, test, and measurement
    • Why do diagnosis
    • Scan chain diagnosis
    • Critical path tracing
    • Algorithmic IC diagnosis
    • Redundancy/Repair (memory and logic)
  3. Variability
    • Design margin versus characterization
    • Statistical behavior (timing models, path-based, block-based)
    • Extreme value theory
    • SSTA, xOCV, and related concepts
    • Test for small delay defects
  4. Putting it all together
    • Reliability issues (repair, aging, burn-in, materials)
    • Relationship of diagnosis, test, and variability
    • Areas for future research