First-Principles Study Of Defect Physics For ZnO And Phase-Change Mechanism For GeSbTe

Semiconductors, considered as one of the critical basements for electronic industry, always push the social development. To some extent, she is also changing the style of our usual life. For her better services for us, we need to fully comprehend some of its technology-related physical problems, including electronic band structure, doping project and responding ability under external field. In our dissertation we employ First-principles-calculation method which is distinct from the traditional experiment method to explore the physics of semiconductor. Here we mainly focuse on two widely-applied semiconductors in microelectronics or optoelectronics, Zinc Oxide (ZnO) and Germanium Antimony Tellurium (GeSbTe).Wide band-gap ZnO attracts many attentions due to its superior performance on optoelectronics. As a kind of practical semicondutor for devices technology it must be made both with n- and p-type, yet the lack of high-quality p-type ZnO (native ZnO with n type) is considered as one of the most intractable problems in defect physics. Hence the application of ZnO must need the breakthrough of the doping problem. Another semiconductor concerned is phase-change material, such as GeSbTe (which is one of the best candidates). Due to its fast transition between the crystalline and amorphous state by laser/electrical pulse, GeSbTe has been widely used in rewritable optical disks and will be the critical candidate for the next generation non-volatile electronic memory. However its phase-change mechanism is still not clear. In our dissertation we employ first-principles calculation to investigate the two critical problems and have been got the four new results.(1) Hydrogen in ZnO. Since Van de Walle in 2000 proposed hydrogen is the possible candidate for native n-type doping, a series of experiments focus the problem of hydrogen in ZnO. However the binding site for hydrogen, one of the most fundamental problems for defect, is still under debate. Lavrov et al found that Hydrogen in Eagle-Picher may stay in the bond center of Zn-O bond parallel to c axis (BC||) with the vibrational mode of 3611cm-1,yet McClusky et al found that in Cermet samples hydrogen stay in the anti-bonding site of oxygen which is vertical to c axis (ABO?) with the vibrational mode of 3326cm-1. A Further investigation observed that there are considerable amount of Ca in Cermet samples. Based on these finding, we investigate the binding site of hydrogen and their stability systematically with first-principles calculation. We find that BC|| is the most stable site for hydrogen without Ca, which is the same situation as Eagle-Picher samples. However, when Ca impurity (replacing Zn) exists in ZnO, the most stable site for Hydrogen is Ca-ABO? site. We employ charge transfer mechanism to explain the phenomena. This finding would be a good candidate of solving the current puzzle for hydrogen binding site in ZnO.(2) The neutral binding effect in ionic semiconductor. In the work (1), we unexpectedly found that the isovalent impurity Ca has very strong binding with H+ (0.7 eV), which is comparable to the binding effect of negative acceptor-positive donor pair or neutral but large distortion defect pair. As a further investigation, we systematically investigate series of semiconductors (includingⅡ-Ⅵ,Ⅲ-Ⅴ,Ⅳ). Besides for Ca in ZnO, we found Ca in BeO, Y in GaN, Y in BN, Al in BN also have strong binding effect with H+, and then we find the following rules: (a) Oxides and Nitrides have tendency of strong neutral binding effect. (b) Sulfides have relative weak binding effect. (c) As the same semiconductor, the smaller electronegativity of impurity (replacing cation) is, the larger the neutral binding is. (d) As the same impurity cation, the smaller iconicity of the semiconductor is, the larger neutral binding is. Based on the charge transferring mechanism, we further proved that the double isovalent impurities (Ca_Zn and O_S) in ZnS have strong binding effect in semiconductor. In all, our results generalize the neutral binding in ionic semiconductor. In fact how the impurity influence the local electrical undulation hold the key defect physics in ionic semiconductor system which is quite different from the case of traditional covalent semiconductor.(3) Electronic Structure for GeSbTe. GeSbTe is one of the best-performance candidates for phase-change material. Generally, its storage process is realized by reversible transition between its amorphous and metastable crystalline state. As a first try of understanding GeSbTe, we investigate the electronic structure from the current popular amorphous model (Kolobov model) and meta-stable model (Rock salt model). We find that (a) essentially both the amorphous and crystalline state keep the p-orbit bonding characteristic in GeSbTe system in stead of any significant sp3 bonding. (b) The large reflectivity contrast stem from the difference of electronic DOS near valence band maximum. (c) Furthermore partial DOS for different element show that there is significant different between amorphous and crystalline state, that is the anti-bonding state near the VBM is obvious for meta-stable crystalline state but not for amorphous state, especially for Ge atom. The previous study on electronic structure for amorphous and crystalline GeSbTe laid the first stone for the next investigation of phase change mechanism with molecular dynamics.(4) Electronic excitation induced solid amorphization. The amorphization under laser/electrical pulse is the―data-writing process‖for storage. Traditional understanding suggests that the amorphization is a melting-quenching process. They simply regard the laser/electrical pulse as heating sources. However, the laser induced electron-hole plasma (EHP) is inevitable, especially the pulse as short as sub-picosecond or several picoseconds. This effect is always neglected for people. With the first-principles molecular dynamics (AIMD), we first demonstrate there is a kind of solid amorphization for GeSbTe system, i.e. it direct go to the amorphous stage instead of experience any long-range atomic diffusion stage. This solid amorphization is induced by the excitation of large amount of electron from its valent band (e.g. 9%). Go in deep with the observation of phase change, we found that the behavior of amorphization is element dependent. At the beginning of phase change, we found the vacancy neighbouring Te frame is collapsed to the Te-Te bonds. From the bond-angle-distribution analysis, we could see that the bond angle of Ge atom (~1000) is most distinct with rock-salt characteristic (~900) bond angle, yet Te and Sb essentially keep bond angle as the crystalline state. This phenomenon of amorphization stem from the unique electronic structure in rock-salt crystalline state. By the charge plot in different energy window, at near VBM (0 ~ -0.66 eV) the charge distributes mainly on the anti-bonding state for Ge and Te and at deeper region for the valence band (-3 ~ -4 eV), the charge for the Sb and Te covalent bonding state is significantly stronger than the one of Ge and Te bonding state. That explain why Ge is most disturbed under electronic excitation. The original Te neighbor distribution after amorphization further support the unique electronic structure hold a key role in this solid phase change process. This finding extends the understanding of GeSbTe system and also displays a possible ultra-fast memory technology in future.In sum, we have shown that several physical problems about electronic science were investigated by first-principles calculation. From these modeling we could obverse the material related phenomenon on atomic scale directly and also some advisable results were obtained. We believe this method will be another good choice for research in electronic filed besides for experiment.