Magneto-Electronic Phase Separation at Perovskite Oxide Interfaces: Origin and Consequences.

摘要

The remarkable functionality of perovskite oxides, when combined with the favorable matching of lattice, chemistry, and thermodynamics at their interfaces, provides many opportunities for new physics and applications. These applications include solid oxide fuel cells, ferroelectric memory, and many other oxide electronic/spintronic devices. The interface between a ferromagnetic (FM) or metallic doped perovskite (e.g. La1-xSrxMnO 3, La1-xSrxCoO3) and a non-magnetic insulator (I) (e.g. SrTiO3, LaAlO3) is a fundamental building block in such structures. The wide range of ground states that give rise to such diverse functionalities in these systems occurs due to close competition between the various degrees of freedom. This close competition however, also poses a serious problem. Maintaining high spin polarization, magnetization, and metallic conductivity at such interfaces has in fact emerged as a significant challenge. This is clearly evident in manganite based magnetic tunnel junctions, where the TMR ratio falls rapidly with temperature.;In this work, we have used the SrTiO3(001)/La1-xSr xCoO3 system to understand the underlying physics of the suppression in magnetic and transport properties near the interface. We have found that this deterioration is a result of nanoscopic magneto-electronic phase separation (MEPS) in the interface region, at compositions that are electronically homogeneous in bulk. The system forms nanoscopic ferromagnetic (FM) clusters embedded in an insulating non-FM matrix near the interface, resulting in suppressed magnetization and insulating transport. Indirect evidence of the interfacial MEPS is made by magnetometry and magnetotransport measurements. Direct proof of the existence of the short-ranged FM clusters is made by small angle neutron scattering (SANS). The thickness (t*) of this magnetic phase separated interfacial region increases remarkably with the decrease in doping (x), finally diverging as the bulk critical doping (xc = 0.175) for metal-insulator-transition (MIT) is approached. Using Z-contrast STEM/EELS measurements we found that this magnetic phase separation is solely chemical in origin, and is driven by subtle depth-wise variations in Sr and O content, leading to the depletion in local hole doping near the interface.;In sharp contrast, LaAlO3(001)/La1-xSrxCoO 3 and SrTiO3(110)/La1-xSrxCoO 3 interfaces display ferromagnetic and metallic behavior to much lower thickness. STEM/EELS measurement for SrTiO3(110)/La1-xSr xCoO3 interfaces reveals a uniform depth-wise Sr and O distribution unlike SrTiO3(001)/La1-xSrxCoO3. We believe that this contrast in the chemical profiles for SrTiO3(110)/La 1-xSrxCoO3 and SrTiO3(001)/La 1-xSrxCoO3 interfaces is a combined effect of surface energy, dopant solubility, and strain state. The findings of this thesis work unequivocally suggest that SrTiO3(110) are better surfaces than SrTiO3(001) for the fabrication of magnetic perovskite based electronic devices.