Front cover image for Study of amorphous metals for nano-scale electronic devices

Study of amorphous metals for nano-scale electronic devices

As CMOS technology continues to scale down, metal gate electrodes are replacing doped polycrystalline silicon gates to eliminate several deleterious effects. However, the use of metal gates presents another problem, namely work function variability ([sigma] GWF), which contributes to the overall threshold voltage variability ([sigma] VT)). [sigma] GWF arises due to the dependence of work function on grain orientation for polycrystalline gate materials. Amorphous metal gates have the potential to eliminate [sigma] GWF. In addition, amorphous metal can be used as a diffusion barrier for copper interconnects, as an electrode for memory devices, and a corrosion resistant coating. This research examines the structural and electrical properties of a variety of amorphous metal alloys, including TaWSiC, TaNi, RuTiN, CoTiN and TiHfN. We present experimental and theoretical discussions on growth mechanism and crystal structure of sputtered and atomic layer deposited films. In the first part, I focus on sputtered amorphous metal gate systems. For the TaWSiC alloy system, I focus on how the composition affects the crystal structure, thermal stability, and electrical properties. It is concluded that both the Ta-to-W ratio and the percentage of Si and C play important roles in stabilizing the amorphous structure. The films with same amount of Ta and W are the most thermally stable, up to 1100°C annealing. At least 5at% Si and C are required to stabilize the amorphous structure for annealing above 400°C. By changing the percentage of each component, a work function tunability of 0.4 eV is achieved in metal/HfO2/SiO2/Si MOSCap structures. For the TaNi alloy system, I focus on how co-deposition and multilayer structure affects the crystal structure. It is found that a wide range of composition of the alloy is amorphous as deposited and are stable above 400°C anneal. For multilayer Ta/Ni structures, samples with individual layer thickness of 0.12nm and 1.2nm are amorphous as-deposited due to intermixing during deposition, and stay amorphous until annealed at 500°C. It is also found that the work function values of the alloy stay close to the Ta work function value of 4.6eV for a large range of compositions due to Ta segregation at the metal oxide interface, which is confirmed by XPS depth profile. In the second part, I focus on the development of ALD titanium nitride-based amorphous gates. The process development and challenges of TiN-based amorphous metal gates are discussed for Ru/CoTiN and TiHfN systems. For Ru/CoTiN, the technical challenges and equipment limitations of Ru and Co ALD deposition in our facility is explored. It is concluded that the Fiji ALD system in Stanford Nano-fabrication Facility (SNF) is not capable of depositing Ru or Co using a non-oxygen chemistry. This limits co-deposition of Ru or Co with TiN, since TiN oxides easily when exposed to oxygen. For the TiHfN alloy system, the crystal structure as a function of the cycle number for each constituent is investigated. It is observed that the columnar growth mechanism of ALD TiN can be either enhanced or interrupted by doping with HfN. Typical grain size for PE-ALD TiN is 5-6nm. By adding HfN, the grain size can vary from near-amorphous (1nm), to up to 10nm. This is attributed to the different nucleation and growth mechanisms of TDMA-Hf on TiN sites, or vice versa. The crystal growth mechanism of the alloy is discussed
Thesis, Dissertation, English, 2015
Stanford University
Academic theses
1 online resource
Submitted to the Department of Materials Science and Engineering