First-principles Study Of SnO₂-based Ferromagnetic Semiconductors

Spintronics aims to combine both spin and charge degrees of electrons as the carrier of information. The development of spintronics significantly enriches the condensed matter physics. The material with good prospects to realize the spintronics practice is ferromagnetic semiconductors. So far, research on diluted magnetic semiconductor is one of the hot topics in the world from the view of both basic research and device application. Searching for the semiconductor with high Curie temperature is one of the important points. Oxide semiconductors with wide gap are the host compounds which can realize the high Curie temperature ferromagnetism. SnO₂ is an important oxide semiconductor with a wide band gap. Its excellent optical transparency, metal-like conductivity, and high chemical stability make it a highly multifunctional material with widespread applicabilities. High temperature ferromagnetism was reported in transition metal doped SnO₂ and undoped SnO₂. It was found that the ferromagnetic properties, such as Curie temperature and magnetic moment, are sensitive to the sample preparation methods and conditions. Theoretically, several computational investigations have been reported on the magnetic properties of SnO₂ base semiconductor. However, a consensus on the origin of thhe ferromagnetism in SnO₂ has not yet been reached. Due to the complexity of diluted magnetic semiconductors and limitations of models based on the mean-field theory and obtained by performing first-principles calculations.In this dissertation we study the electronic structure and magnetic properties based on SnO₂ host by using first-principles density functional method to get the origin of ferromagnetism. We discussed the effect of dopants, defects and low dimension on the magnetism. All the results of this dissertation have been calculated by the WIEN2k, which is based on full-potential and linearized augmented plane wave basis set. The results are as follows: The effects of Co dopants and oxygen vacancies on the electronic structure and magnetic properties of the Co-doped SnO₂ are studied with GGA and GGA+U. The Co atoms favorably substitute on neighboring sites of the metal sublattice. Without oxygen vacancies, the Co atoms are at low spin state, which independent on concentration and distribution of Co atoms, and only the magnetic coupling between nearest-neighbor Co atoms is ferromagnetic. Oxygen vacancies tend to locate near the Co atoms. Their presence strongly increases the local magnetic moments of Co atoms, which depend sensitively on the concentration and distribution of Co atoms. Moreover, oxygen vacancies can induce the long-range ferromagnetic coupling between well-separated Co atoms through the spin-split impurity band exchange mechanism. The addition of effective U Co transforms the ground state of Co-doped SnO₂ to insulating from half-metallic and the coupling between nearest neighbor Co spins to weak antimagnetic from strong ferromagnetic. GGA+ U Co calculations show that the pure substitutional Co defects in SnO₂ can not induce the room-temperature ferromagnetism. Oxygen vacancies tend to locate near Co atoms. Their presence increases the magnetic moment of Co and induces the FM coupling between two Co spins with large Co-Co distance. According to our calculated results oxygen vacancies enhancing the room temperature ferromagnetism observed experimentally can be deduced from three aspects: enlarge the spin-split of Co-3d states; induce a spin-split impurity band, which hybridizes with the Co-3d states at EF; induce charge transfer from impurity states to Co-3d states. By more charge transfer and larger spin split of Co-3d and impurity states, the addition of U Co enhances the ferromagnetic stability of the system with oxygen vacancies and leads the high Curie temperature. These theoretical results testify to the conclusions of the spin-split impurity band model. By applying a Coulomb U O on O 2s orbital, the band gap is corrected for all calculations to study the impact of band-gap underestimation on the ferromagnetic mechanism. The conclusions derived from GGA + UCo calculations are not changed by the correction of band gap.For Ni doped SnO₂, Ni itself can not induce magnetism without oxygen vacancies. This is consistent with the experimental results that the ferromagnetism of Ni doped SnO₂ films disappeared after annealed in oxygen rich atmosphere. Oxygen vacancies tend to locate near Ni atoms. Their presence increases the magnetic moment of Ni to 1.4μB/Ni from 0. The oxygen atoms located inside the octahedron of Ni are spin polarized more or less, but the farther Sn and oxygen are not polarized. For that the locate magnetic moments can hardly long-range couple. Calculated results also show that there is nearly no coupling between the farthest Ni ions. The coupling is stronger for Ni ions with nearer distance, but the couple is antiferromagnetic. The addition of U Nimake the couples between Ni transition towards ferromagnetism. The intensity of ferromagnetic coupling between two farthest Ni ions becomes double. However the band-gap correction makes the couples towards antiferromagnetism. From the calculated results we expect the couple between Ni ions flip with the different distance between them. Sure the couple between two Ni ions can be ferromagnetic at some distances.The effect of tin interstitial Sni and oxygen vacancy VO on the electronic structure and magnetic properties of undoped SnO₂ is investigated by means of density functional calculations. Only single positively charged O vacancies VO1+ can induce local moments in bulk SnO₂. Self-consistent band gap correction, which is achieved by adding a Coulomb U on O-2s orbital, leads the spin-up gap level of Vo1+ to be fully filled from the partially occupation calculated by GGA. Consequently, the coupling between two Vo1+ becomes antiferromagnetic from ferromagnetic. So a self-consistent band gap correction is essential for the correct description of magnetism in wide-gap SnO₂. These results indicate that O vacancy in bulk SnO₂ can not induce ferromagnetism, while the atoms or defects located at the surface or substrate interface should play a key role in turning the ferromagnetism observed in undoped SnO₂.The effect of surface atom, defect and defect complex in the most (110) surface of SnO₂ on the electronic structure and magnetic properties of (110) surface is investigated. The results show the surface terminated with 1-fold coordinated O is magnetic, the main contributor is the 1-fold coordinated O while the 2-fold coordinated O is non spin polarized. Though it is inspiring to find the magnetic atom, but it is inconsistent with the experimental phenomenon. Experiment reported that ferromagnetism appears under O poor atmosphere in undoped SnO₂, however, the 1-fold coordinated O atoms may exist only under O rich atmosphere. So 1-fold coordinated O is not the origin of magnetism observe in undoped SnO₂. Single bridging or in-plane oxygen vacancy in stoichiometric and reduced (110) surface of SnO₂, just as the single oxygen vacancy in bulk SnO₂, can not induce magnetic moment. Moreover, we studied the effect of O vacancy complex on the magnetic properties and found that the in-plane O vacancy complex at reduced (110) surface can induce magnetic moment while the complex can not induce magnetism in bulk SnO₂. So we conclude that the ferromagnetism in undoped SnO₂ origins from the O vacancy complex at surfaces.In conclude, except the dopants, the oxygen vacancy plays an important role in the ferromagnetism of Co and Ni doped SnO₂. For undoped SnO₂, both defects and low-dimension play the key role in the ferromagnetism.