Search

Evaluation of diffusion barrier integrity is an important issue in advanced interconnects. A diffusion barrier separating Cu from low-k must be as thin as possible and must not contain pinholes. We have developed a method for measuring pinhole density in diffusion barriers deposited on low-k materials. The method employs ellipsometric porosimetry for measuring diffusion of toluene in a porous low-k film beneath the barrier in question. [Copyright &y& Elsevier]

Implementation of Cu/low-k in advanced interconnections requires a diffusion barrier to prevent copper penetration in low-k dielectrics. The barrier should be continuous to prevent copper diffusion and thin enough to keep interconnection line resistance low. Deposition of a diffusion barrier becomes an issue when porous low-k dielectrics are used. We developed a Monte Carlo simulation model to describe deposition of a diffusion barrier on a porous low-k film. The model provided explanation for the sealing behavior of different porous film by TaN diffusion barrier. Previously we have shown that TaN barrier integrity depends on chemical nature of the substrate rather than on porous structure: the same barrier can be continuous on oxycarbide (SiOCH) but non-continuous on HSQ (hydrogen silsesquioxane) although porous structures of the two films are similar. Using the model, we show that surface diffusion of TaN during deposition plays a key role in continuous barrier formation. Presence of carbon suppresses TaN diffusion (probably by TaC formation) and the barrier does not penetrate deep into the film forming a continuous layer on the top surface. The model is also able to predict sealing behavior for different porous low-k films with different porosity/pore sizes. [Copyright &y& Elsevier]

Galvanic corrosion of copper as a function of barrier metal type together with the effect of acid type is investigated. Refractory metals such as Ta, Ti, and W are used as the barrier metals. HNO3-based and H2O2-based aqueous solutions are used to analyze the effect of acid type of the CMP slurry on the galvanic corrosion. The results show that the galvanic etch rate of copper is affected by the barrier metal type. In addition, the type of solution used is found to influence the barrier metal as well as the copper. Especially for the case of W, the change in the direction of the electrical current flow between copper and the barrier metal depending on the type of solution used seems to alter the galvanic reaction mechanism. [Copyright &y& Elsevier]