Knowledge Center
Knowledge Center

Blech Effect

A reverse force to electromigration.


The Blech effect works in opposition to electromigration. Migration creates tensile stress at the upstream, cathode, end of the line, and compressive stress at the downstream, anode end. The sidewalls, being rigid, supply back pressure against this stress, slowing migration. The Blech effect is line length dependent: shorter lines offer more resistance. S. Thrasher and colleagues at Motorola found that there is a critical value of the length-current density product below which no migration occurs.

This has led to electromigration-driven design rules, limiting the allowable wire length as a function of current density. It is not always clear, however, just where line segments begin and end for the purpose of evaluating electromigration risk, and it is not always easy to quantify the benefits that the Blech effect can provide. Researchers at IBM conducted an extensive analysis to determine the best ways to define the length of non-trivial line segments, and to evaluate the results.

The difficulty in balancing resistance, gap fill, and electromigration becomes especially clear when considering proposals for alternative barrier and seed layers. Thinner barrier layers are desirable because they increase the portion of the total line width available for copper. However, in addition to the risk that a thinner diffusion barrier will be less effective, thinner layers are generally less rigid. A thinner layer might reduce or eliminate the Blech effect back pressure against electromigration, which the designer may be depending on to keep reliability within specifications. Here, paradoxically, longer lines actually have an advantage over shorter lines. Longer lines do not benefit from the Blech effect, and therefore their performance does not suffer if it is reduced. An alternative barrier might also have weaker copper adhesion, increasing the potential for migration along trench sidewalls.

The effect is named after Ilan Blech of Israel’s Technion, who in the 1970s discovered that when current was passed through gold on a substrate the upstream side moved with the current flow but the other side remained fixed. But if the distance was short enough, there would be no electromigration.