First Principle Studies On Several Molecular Magnets And Magnetic Strong-Correlation Systems

With the development of computer, the computation of electronic structure for practical materials (up to one hundred atoms) has become available. On the other hand, density function theory gives a firmly theoretical foundation for the study of ground states of materials. Then the first principle calculation based on density function theory has become one of the most powerful tools in condensed matters. With the first principle calculation, we can not only get the ground state properties of material, but also simulate their behaviors under different conditions. This means that we can predict the properties of materials or design the required materials. Here we use the first principle calculation based on density function theory to study the electronic structures of several magnetic materials. Our work can be dived into two parts: one is about the moment distributions and magnetic couplings in molecular magnets, the other is about the ground state properties of two strongly correlated systems with element Ce. The calculations have been done with the WIEN2K code. It adopts the full-potential linearized augmented plane wave method, which is among the most accurate methods for performing electronic structure calculations for crystals. Traditional magnetic materials are metals with 3d or 4f electrons. But they can not fulfill the requirements under some conditions with the development of science and technology. So it is important to find new magnetic materials. During the last two decades, people have fabricated a lot of organic magnets or metal-organic magnets (molecular magnet). The molecular magnet can be dived into different parts corresponding to the different types of the constitutive magnetic centres: purely organic magnet, inorganic coordination complex and charge transfer salt. In this paper, we first study the electronic structure and magnetic coupling of purely organic ferromagnet α-HQNN. α-HQNN is the first hydrogen-bonded organic ferromagnet based on the nitronyl nitroxide radical. Low-temperature experiment showed that it became ferromagnetic at about 0.5K, but the spin moment distribution in the molecule has not been measured. M. M. Matsushita et. al. believed that the spin moment distribution in α-HQNN should be similar with that in PhNN, then they proposed that the ferromagnetic couplings between α-HQNN molecules are due to the spin polarization through the intermolecular or intramolecular hydrogen bond. From the calculation, we get the real spin distribution in α-HQNN molecule. But we find that there is large difference between our result and that assumed by M. M. Matsushita et. al. Our result indicated that the ferromagnetic couplings between α-HQNN molecules should be attributed to the spin delocalization effect, not the spin polarization effect. Second, we study the magnetic couplings in azido-metal complexes. Azido is one kind of ligand that is composed of three nitrogen atoms. It can bridge metal ions with μ-1,1,μ-1,3,or μ-1,1,3 modes. There are two viewpoints about the magnetic coupling between two metal ions through azido ligand: one proposes that it is due to the spin polarization of azido ligand, while the other proposes that it is due to the spin delocalization from magnetic ions to azido ligands. The first is simple and clear, and can explain many early experiments, so it is often used to explain the magnetic coupling for azido-metal complexes. We select several typical azido-metal complexes, and calculate their electronic structures. The systemic studies show that the magnetic couplings between two metal ions through azido ligands are due to the spin delocalization from metal ions to azido ligand, ,not spin polarization, irrespectively of the bridging modes. But in some cases, the spin polarization also has weak contribution to the magnetic coupling. The exchange and correlation interaction in the density function theory are often approximated by LD(S)A or GGA. The validity of LD(S)A or GGA has been proved by many first principle calculations. But there are problems when we use LD(S)A or GGA to calculate strongly correlated system, such as materials with 3d transition metal or 4f rare earth metals. This is due to the fact in LD(S)A or GGA, the correlated effect of electrons in the narrow 3d or 4f bands are described the mean field manner, which is about onemagnitude smaller than the reality. In order to overcome this problem, the so called "L(S)DA+U"method has been proposed by people recently. In this paper, we adopt this method to calculate the electronic structure of two kinds of Ce-derived compounds: CeAg和CeRu2Ge2,to see if the strongly correlated interaction affect the properties of ground state. Our results reveal that cubic-tetragonal structural transition of CeAg does not affect by the strongly correlated effect, while the magnetic moment and anisotropy of CeRu2Ge2 are affected largely by the strongly correlated effect.