Cobalt and Ruthenium Interconnect Metallization is the adoption of alternative conductor metals to replace copper in the narrowest BEOL interconnect levels, where the effective resistivity of copper degrades dramatically due to electron scattering at grain boundaries and interfaces, making cobalt (Co) and ruthenium (Ru) increasingly attractive options despite their higher bulk resistivity — driven by the crossover point where copper's practical resistance in nanoscale wires exceeds that of metals with superior scaling behavior.

• Copper Scaling Problem: Copper's bulk resistivity of 1.7 micro-ohm-cm is the lowest among practical interconnect metals, but at line widths below 15-20 nm, electron mean free path scattering at grain boundaries and barrier interfaces causes the effective resistivity to increase by 3-5 times; additionally, the required TaN/Ta barrier and Cu seed layers consume an increasing fraction of the wire cross-section, further reducing the effective conducting area.
• Cobalt Advantages: Cobalt has a shorter electron mean free path (approximately 8 nm versus 39 nm for copper), meaning its resistivity scales more gracefully at narrow dimensions; cobalt can be deposited by CVD with excellent conformality and does not require a thick diffusion barrier because cobalt itself has lower diffusivity in dielectrics than copper.
• Cobalt Integration: Cobalt interconnects at the M0 and M1 levels use a thin TiN liner of 1-2 nm for adhesion, followed by CVD cobalt fill using Co2(CO)8 or similar precursors; CMP removes overburden metal, and a dielectric cap provides oxidation protection; cobalt's lower electromigration activation energy requires careful current density limits.
• Ruthenium Advantages: Ruthenium has a bulk resistivity of 7.1 micro-ohm-cm and an electron mean free path of approximately 6 nm, providing even better resistivity scaling than cobalt at the smallest dimensions; ruthenium also does not require a diffusion barrier when integrated with certain low-k dielectrics, enabling a barrier-less integration scheme that maximizes the conducting cross-section.
• Barrier-Less Integration: Ruthenium's chemical stability and low diffusivity into SiO2-based dielectrics allow direct metal deposition without a barrier layer; this eliminates the 2-3 nm of cross-section consumed by traditional TaN/Ta barriers, recovering 30-50 percent of the conducting area at sub-10 nm line widths.
• Deposition Techniques: ALD and CVD ruthenium deposition using RuO4 or Ru(EtCp)2 precursors achieves conformal, void-free fill of high-aspect-ratio damascene trenches; selective deposition on metal versus dielectric surfaces is also being developed to enable bottom-up fill without seed layers.
• Subtractive Patterning: Unlike copper, which must use damascene processing because it cannot be easily dry-etched, both cobalt and ruthenium can be patterned by subtractive (deposit-and-etch) methods using chlorine or oxygen-based plasma chemistries; subtractive patterning eliminates CMP dishing and erosion issues and simplifies the process flow.
• Hybrid Metallization: Production BEOL stacks may use cobalt or ruthenium for the tightest-pitch local interconnect levels (M0-M2) while retaining copper for wider semi-global and global levels where copper's lower bulk resistivity still provides an advantage. The transition to cobalt and ruthenium interconnects represents a fundamental materials change in semiconductor manufacturing, driven by the physical reality that copper's scaling limitations make alternative metals essential for continued interconnect performance improvement.