• DocumentCode
    1451358
  • Title

    Breaking with tradition in mathematical metal reliability modeling

  • Author

    Borucki, Len

  • Author_Institution
    Predictive Eng. Lab., Motorola Inc., Mesa, AZ, USA
  • Volume
    4
  • Issue
    3
  • fYear
    1997
  • Firstpage
    10
  • Lastpage
    12
  • Abstract
    One of the limiting factors in designing advanced microelectronics is the size and complexity of the wiring used to connect electronic components on a chip. Although transistors can be scaled to smaller sizes with well-understood rules, metallization cannot. Typically, scaling raises the current density as the lines themselves are made thinner. This increases the chance of circuit failure due to electromigration. Another kind of failure that occurs is stress voiding, in which the stresses due to thermal-expansion mismatches of nearby materials and the electromigration of atoms actually break a line. The Motorola Predictive Engineering Laboratory has been developing a computer model for stress and electromigration that includes considerably more physics than the routinely-used statistical models. If we can project metal lifetimes more accurately, we could generate better design rules and thus produce less expensive products that are less limited in area or performance by metallization. We are working directly with engineers who formulate and qualify our metallization processes and extract rules for product designers company-wide. The goal is to affect PowerPC and SmartPower designs
  • Keywords
    circuit analysis computing; digital simulation; electromigration; integrated circuit design; integrated circuit metallisation; integrated circuit modelling; integrated circuit reliability; metals; reliability theory; thermal stresses; wire-wrap connections; Motorola Predictive Engineering Laboratory; PowerPC; SmartPower; advanced microelectronics design; circuit failure; computer model; current density; design rules; electromigration; mathematical metal reliability modeling; metal lifetimes; metallization; performance; scaling; stress voiding; thermal-expansion mismatches; wiring; Current density; Electromigration; Electronic components; Mathematical model; Metallization; Microelectronics; Predictive models; Thermal stresses; Transistors; Wiring;
  • fLanguage
    English
  • Journal_Title
    Computational Science & Engineering, IEEE
  • Publisher
    ieee
  • ISSN
    1070-9924
  • Type

    jour

  • DOI
    10.1109/99.615425
  • Filename
    615425