A Biased Resistor Network Model for Electromigration Failure and Related Phenomena in Metallic Lines

TL;DR

This paper introduces a biased resistor network model to simulate electromigration failure in metallic lines, capturing void formation, stress growth, temperature effects, and alloying, with results matching experimental data.

Contribution

The study presents a comprehensive resistor network model that unifies various electromigration phenomena and accurately predicts failure times and damage patterns.

Findings

Model agrees with experimental TTF distributions.

Accurately reproduces damage patterns and resistance evolution.

Describes electromigration saturation and noise properties.

Abstract

Electromigration phenomena in metallic lines are studied by using a biased resistor network model. The void formation induced by the electron wind is simulated by a stochastic process of resistor breaking, while the growth of mechanical stress inside the line is described by an antagonist process of recovery of the broken resistors. The model accounts for the existence of temperature gradients due to current crowding and Joule heating. Alloying effects are also accounted for. Monte Carlo simulations allow the study within a unified theoretical framework of a variety of relevant features related to the electromigration. The predictions of the model are in excellent agreement with the experiments and in particular with the degradation towards electrical breakdown of stressed Al-Cu thin metallic lines. Detailed investigations refer to the damage pattern, the distribution of the times to failure (TTFs), the generalized Black's law, the time evolution of the resistance, including the early-stage change due to alloying effects and the electromigration saturation appearing at low current densities or for short line lengths. The dependence of the TTFs on the length and width of the metallic line is also well reproduced. Finally, the model successfully describes the resistance noise properties under steady state conditions.