By S. Rossnagel, IBM Research Div., T.J. Watson Research Center

The original interest in ruthenium (Ru) was based on the ability to directly plate copper (Cu) onto Ru without a Cu seed layer. There are perhaps five or six other materials that could also be used in this manner (Pt, Ir, Re, etc.), but Ru had other advantages. Initial work suggested that the Cu-Ru interface had low electromigration (EM), which is a surface-driven effect in Cu, and the adhesion of Cu to Ru appears good. The solubility is low, so the net effect on Cu resistivity (due to impurity doping) is minimal.

Ru was also compatible with CVD and ALD technology. Such compatibility is a major constraint with Ta technology, where the nitride (TaN) can be fabricated using ALD with metalorganic precursors, but Ta (metal) cannot be fabricated without the use of halogen precursors (e.g. TaCl₅), which are corrosion risks for Cu technology. Recent work with ALD of Ru has been very encouraging and a number of metalorganic Ru precursors are available.