Abstract
Initially, the present dissertation deals with the preparation and characterization of metal-germanium junctions. The research on these metal-germanium junctions is mainly important for the two following reasons: firstly, because through this research, fundamental characteristics as the germanium charge neutrality level can be studied and secondly because these junctions can be used for the source and drain formation in germanium MOSFETs replacing the p-n junctions. Indeed, germanium p-n junctions present a big challenge because dopants in Ge present low solid solubility, incomplete activation and enhanced diffusion. The Schottky barrier height in metal/Ge contacts is only weakly dependent on the metal work function indicating strong Fermi- level pinning close to the Bardeen limit. The pinning factor S is about 0.05 and the charge neutrality level (CNL) is only about 0.09 eV above the top of the valence band. Because of this, the Fermi level lies higher than CNL in most cases of intere ...
Initially, the present dissertation deals with the preparation and characterization of metal-germanium junctions. The research on these metal-germanium junctions is mainly important for the two following reasons: firstly, because through this research, fundamental characteristics as the germanium charge neutrality level can be studied and secondly because these junctions can be used for the source and drain formation in germanium MOSFETs replacing the p-n junctions. Indeed, germanium p-n junctions present a big challenge because dopants in Ge present low solid solubility, incomplete activation and enhanced diffusion. The Schottky barrier height in metal/Ge contacts is only weakly dependent on the metal work function indicating strong Fermi- level pinning close to the Bardeen limit. The pinning factor S is about 0.05 and the charge neutrality level (CNL) is only about 0.09 eV above the top of the valence band. Because of this, the Fermi level lies higher than CNL in most cases of interest so that unpassivated acceptor-like gap states at the interface are easily filled, building-up a net negative fixed charge. This could prevent efficient inversion of a p-type Ge surface in a metal-oxide-semiconductor structure. Moreover, this dissertation deals with the modelling of negatively charged states of germanium surfaces and interfaces. This modelling is useful for the understanding of various problems which relate with the observed poor performance of Ge n-MOSFETs (despite of the improvement in germanium surface passivation) and the observed undesired positive threshold voltage shift of Ge p-MOSFETs (this voltage is expected to be negative). Modeling based on surface charge neutrality predicts that the Ge surface tends to be p-type, irrespective of the bulk conductivity. This is a consequence of the fact that the Ge band gap is small and the charge neutrality level lies low in the gap very close to the valence band, probably determined by low-lying, unpassivated surface dangling bond acceptors or other defects. According to the model, the acceptor defects build negative charge, inverting the surface of n-type Ge at no gate bias for low doping concentration (<10¹⁶ cm⁻³) and moderate or high interface state densities (>5x10¹¹ eV⁻¹ cm⁻²). This is predicted to cause undesired positive threshold voltage shift in the range of 0.2 to 0.4 V in Ge p- channel field effect transistors. The model also predicts that inversion in n-channel field effect transistors is inhibited, which could be related to the observed poor performance of these devices. Finally, this dissertation deals with the ZrO₂/GeO₂/Ge structure preparation and characterization. Finding an oxide with high dielectric constant along with desirable insulating characteristics (leakage current Jg~1.2x10³ A/cm²) is necessary requirement so as the success criteria of the ITRS specifications for the 22 nm node can be satisfied. Electrical data on ZrO₂/GeO₂ stacks prepared by atomic oxygen beam deposition on Ge at 225°C reveal a relatively weak dependence of the stack equivalent oxide thickness upon the ZrO₂ thickness. This trend points to a very high zirconia dielectric permittivity (k) value which is estimated to be around 44. This is indicative of zirconia crystallization into a tetragonal phase which is also supported by x-ray diffraction data. X-ray photoelectron spectroscopy analysis is in line with the assumption that, due to a finite GeO₂ decomposition, Ge is incorporated into the growing ZrO₂, thus, stabilizing the high-k tetragonal phase. Tetragonal ZrO₂ can be produced, also, with the Ge atoms incorporation during the ZrO₂ growth without being obligatory the GeO₂ existence. A Ge atom percentage of ~6% along with growth temperature of 225°C stabilizes optimum the ZrO₂ tetragonal phase. The stabilization of the ZrO₂ tetragonal phase has two advantages: firstly, low EOT values (because of the high dielectric constant) and secondly, very low leakage currents (high energy gap) which is very important if someone takes into account that the oxide energy gap reduces with the increase of the oxide dielectric constant.
show more
