Optische Eigenschaften von Phasenwechselmaterialien für zukünftige optische und elektronische Speicheranwendungen

摘要

Phase Change Materials offer a unique combination of physical properties, thus they yield successful applications. They are used in re-writable optical data storage and will be in the near future also used in electrical data storage. The alloys, employed up to now in optical data storage are developed empirically. For future applications, especially electrical data storage, it will be important to have design rules to tailor certain physical properties of phase change materials. To find these, a basic knowledge of the characteristic properties of phase change materials is necessary. In this work the optical properties of the amorphous and crystalline phases play a crucial role. Optical properties contain information about the chemical bonding and even about electrical transport parameters. If free carrier have enough influence on the optical properties, the conductivity can be calculated. Additionally, for large Drude relaxation times, this parameter itself, the quotient of the carrier concentration and the effective mass as well as the product of the mobility and the effective mass can be calculated. If the relaxation times are too small, it is still possible to calculate bounds for these transport parameters. The optical properties were measured with FTIR spectroscopy in the infrared and spectroscopic ellipsometry in the visible region in combination with metallic reflectors. The comparison of optical, electrical and structural properties will lead to deeper insight into the physics of phase change materials. The analysis of the polarisability of different phase change materials show that the optical properties of amorphous systems can be described very well depending on the density and the stoichiometry. The polarisability of crystalline systems is remarkable high, thus the chemical bonding must have been changed upon crystallisation. After crystallisation, the increase of middle range order in the system cause the formation of resonant bonds, additionally to covalent bonds, which are ordinary for known semiconductors. Resonant bonds arise, when there are more bonds (six in the rock salt structure) than allowed, following the 8-N rule (three p-electrons). Crystalline phase change materials usually have conductivities over 10 S/cm, free carriers strongly influence the FTIR spectrum. They can be described within the Drude model. Both, the analysis of the optical spectra and electrical measurements result in conductivities of the same order of magnitude. This and the extremely short Drude relaxation times lead to the assumption, that the scattering mechanism cannot be explained by grain boundaries, but a microscopic material property. Carrier concentrations are only few orders of magnitude below those of metals. This is a first hint, that crystalline phase change materials are degenerated semiconductors. With gold as metallic reflector, diffusion and reaction processes were confirmed. This has a marginal influence of the optical properties, thus silicon substrates or aluminum reflectors were used afterwards. Some crystalline phase change materials, for example GeSbTe alloys, show a decrease of the resistivity upon annealing of two orders of magnitude without changing the structure. Other phase change materials, like GeTe, do not show this effect. This important and interesting effect was investigated with different methods. For crystalline phase change materials optical properties and electrical properties show similar dependencies on the annealing temperatures. FTIR spectra of crystalline Ge1Sb2Te4 or Ge2Sb2Te5 show a systematic trend in the Drude term upon annealing, as well as in the interband transitions. Thus, a change of the electrical transport parameters goes along with a change of resonant bonding. Hence, the knowledge of the chemical bonding helps describing electrical transport of phase change materials. In crystalline GeTe both the spectra and the electrical properties are independent of the annealing conditions. Furthermore, amorphous systems were investigated upon annealing. Structural relaxation processes are accelerated upon annealing, because energy barriers have to be overcome and this process is thermally activated. FTIR spectra show an increase of the band gap upon annealing. So the standard transport model provide an opportunity to explain drift. This phenomenon is an increase of the resistivity with time in amorphous phase change materials. Additionally optical measurements at a semi-crystalline GeTe thin film in combination with structural, calorimetric and electrical measurements give information about the heterogeneous crystal growth mechanism. With the aid of a cryostat, FTIR measurements can be performed temperature-dependent in the range of 5-350 K. Because structural changes of the film can be excluded, a purely electronic effect was measured. The temperature-dependency of the bandgap is clearly larger in amorphous systems in comparison to crystalline systems. Among each other these dependencies are rather similar. The knowledge of the temperature-dependency of the bandgap is important for simulations of the temperature-dependent conductivity plus the interpretation of Seebeck- or MPC (modulated photocurrent) data. Furthermore, there is no freeze out of free carriers at 5 K. Materials, where transport parameters could be determined exactly, show temperature-dependencies of the conductivity and other transport parameters similar to metals or degenerated semiconductors, respectively.