Modeling the atomic and electronic structure of Metal-Metal, Metal-Semiconductor and Semiconductor-oxide interfaces

Ganesh K Hegde, Purdue University

Abstract

The continuous downward scaling of electronic devices has renewed attention on the importance of the role of material interfaces in the functioning of key components in electronic technology in recent times. It has also brought into focus the utility of atomistic modeling in providing insights from a materials design perspective. In this thesis, a combination of Semi Empirical Tight-Binding (TB), first-principles Density Functional Theory and Reactive Molecular Dynamics (MD) modeling is used to study aspects of the electronic and atomic structure of three such 'canonical' material interfaces—Metal-Metal, Metal-Semiconductor and Semiconductor oxide interfaces. An important contribution of this thesis is the development of a novel TB model of the electronic structure of industrially relevant metals such as Cu, Ag, Au, Al. The model has been validated for accuracy and transferability against DFT calculations and has been integrated into the industry standard Nano Electronic MOdeling Tool 5 (NEMO5) simulator. In this thesis, the model has been used to provide insight into the role of quantum mechanical confinement, grain orientation in the conductivity degradation of polycrystalline Cu interconnects. The use of homogeneous compressive strain in improving the conductivity of copper has also been studied using this model. The impact of interface chemistry on the electronic structure of Cu-Si interfaces has been investigated using Reactive MD and DFT transport calculations. Strong Fermi Level Pinning (FLP) is seen at these interfaces independent of doping concentrations providing theoretical support for the Metal-Induced Gap States (MIGS) model in describing FLP at Metal-Si interfaces. A DFT reaction pathway approach has been used to provide theoretical insight into experimentally observed anisotropy in III-V—Atomic Layer Deposited (ALD) Al2O3 N-Metal Oxide Semiconductor (NMOS) device performance with changes in device substrate orientation from (111)A to (111)B. The significant difference in eventual interface chemistry on the respective surfaces leads to radically different electronic structures and correlates well with observed experimental anisotropy in device performance.

Degree

Ph.D.

Advisors

Klimeck, Purdue University.

Subject Area

Electrical engineering|Condensed matter physics|Materials science

COinS