Modelling the Atomic Layer Deposition of Cobalt and Ruthenium on NHx-Terminated Metal Surfaces

Abstract

Atomic layer deposition (ALD) is widely used in microelectronics and semiconductor industry to deposit very thin films as part of device fabrication in nano- or subnano-dimensions. The key advantages of ALD are the conformality and precise thickness control at the atomic scale, which are difficult for physical or traditional chemical vapor deposition methods. Cobalt (Co) and Ruthenium (Ru) are used as a seed layers for metallization of interconnects. They are also potential materials for the electrode in dynamic random-access memory (DRAM) capacitors and metal-oxide-semiconductor field-effect transistors (MOSFETs). Plasma-enhanced ALD (PE-ALD) is used for low-temperature thin film growth by alternating exposures of metal precursors and plasma reactants. During the PE-ALD growth of metals, N-plasma, for example, NH3 or a mixture of N2 and H2, has been applied to avoid surface metal oxidation. The PE-ALD of Ru and Co has been experimentally investigated using metal precursors such as RuCp2, Ru(EtCp)2, CoCp2, and CoCp(CO)2. However, the reaction mechanism is not clear and theoretical studies on the reaction mechanism is entirely lacking.In this presentation, we study the PE-ALD growth of Co and Ru by first principle calculations. The (001) surface of both metals, with a hexagonal structure, is the most stable and the (100) surface with a zigzag structure is less stable but has high reactivity. These two surfaces allow the study of the influence of the surface facet. The surface saturation coverage was studied by considering individual adsorption and co-adsorption of NH and NH2 to terminate both surfaces. On the (001) surface, the zero K saturation coverage for NH is 1ML on Ru and 6/9ML on Co. For NH2, the saturation coverages are 6/9ML on Co and Ru surfaces. On the (100) surface, the saturation coverages are 2ML for NH on Ru and Co surfaces, 1.33 ML for NH2 on Ru, and 1ML for NH2 on Co surface. The higher saturation coverage on (100) surface is attributed to the unique trench structure, which provides more available surface sites than that of (001) surface. We also consider co-adsorption of NH and NH2 on (001) and (100) surfaces. The results are then analyzed with ab initio thermodynamics by calculating the Gibbs free energy. Both the ultra-high vacuum (UHV) condition and standard ALD operating condition are used to elucidate the effect of pressure and temperature on the termination of metal surfaces.The adsorption and reactions of metal precursors (CoCp2 and RuCp2) on NHx terminated metal surfaces were investigated with the inclusion of van der Waals corrections. Two possible adsorption structures, namely the precursor lying parallel and perpendicular to the NHx-terminated surfaces, were considered at zero K and ALD conditions. Possible reactions include: precursor adsorption, first hydrogen transfer, first Cp ligand dissociation, first Cp ligand desorption, second hydrogen transfer, and second Cp ligand desorption. The barrier for proton transfer was calculated using climbing image nudged elastic band (CI-NEB) method. Our results show that (100) surface has higher activity than (001) surface. In addition, the Cp ligand elimination of CoCp2 has lower barrier than RuCp2, regardless of surface facet. After the metal precursor pulse, the surface is terminated with MCp fragment or M atom depending on surface facet, where M is Ru or Co. The surface N atom and H atom can be eliminated during the following plasma step by forming NH3, N2 or H2. After a full cycle, the surface is an NHx-terminated metal surface and ready for the next cycle. This work will be important to reveal the mechanism and feasibility of atomic layer deposition of metals using N-plasma.