The resistivity switching of iron doped strontium titanate single crystals and thin films

Abstract

The main objective of this work was to identify and understand the physical mechanisms underlying the phenomenon of the RS in the Fe doped SrTiO3 material. Two basic types of materials were investigated: the single crystals and thin lms. The main reason for this approach was to obtain the information about both: the fundamental physics (single crystals) and the material more suited for the applications (thin lms). The samples with di erent iron concentrations, were modi ed by the temperature, electric eld and chemical gradients. Resulting changes were investigated by a number of techniques that are characteristic to the solid state physics in the macro- and nano- scales. Completed studies have shown that: Crystallographic structure of the starting crystals was in good agreement with the literature. Surprisingly high level of the Fe concentration variation, even for the relatively low Fe doping, was found. Local conductivity in nanoscale showed characteristic conducting stripes, not observed previously. Fe valence state was found to be a mixture of 2+ and 3+, contrary to the most of the reports on similar systems. Moderately high temperature treatment under oxidizing and reducing conditions led to huge chemical modi cation of the crystal surface, including formation of new phases and the decomposition of STO phase. High mobility of the Fe ions above 700 C allowed for signi cant iron migration and preferential atoms localization (close to the extended defects). Especially interesting is that the crystals exposed to the high temperaturę under various conditions undergone transformation and decomposition to other phases. This remains in the contrary to conventional model, where only the oxygen content in the form of SrTiO3􀀀 was changing. Moreover, the a nity of the Fe ions to locate close to the extended defects was shown, as an formation of lament like Fe structure. The nanoscale RS behavior was also evident and did not di er signi cantly from the behavior in the undoped STO crystals. It was concluded, that the basic mechanism of RS in Fe doped STO is not in uenced by the iron doping. Similar results can be observed under the electrical stress. The electroformation and subsequent electro-coloration phenomena were investigated. The most prominent results showed: Electrical stimuli led to the I-M transition. I-M transition was connected to changes in the oxygen vacancy concentration and the movement of the oxigen ions along the extended defect network. Evolution of the color front was documented, and compared to the removal of oxygen ions from the material. Total number of the oxygen ions removed from the material during electroreduction experiment was estimated. RS behavior was found and measured both in macro- and nano-scales. Fast reaction of the electrical behavior to the surrounding atmosphere conrmed, that the oxygen can be removed and introduced to the system quite easily. Strong connection between the evolution of the color front and the electrical properties of the samples were shown. Also the fact that the Fe doped STO crystal is in fact 'open' to the removal and introduction of the oxygen ions from the surrounding atmosphere was found. The investigation of the thin lms yielded promising results, especially in the similarities with the single crystal behavior. The most interesting ndings consist of: Many hints that the Fe ions were inhomogeneously distributed in the STO matrix was found. Fe was found to be very mobile and typically was found to migrate away from the surface into the bulk. High susceptibility of the electrical parameters to the electrical stress led to high variation in the measured activation energy values. The RS behavior was con rmed and no clear voltage threshold for the set and reset voltage was found (up to several mV). This results showed many similarities of the STO thin lm and single crystals behavior. It seems possible to maintain good single crystal properties in more applicative thin lms. What is more, the pursue for the 'perfect' or defect-free thin lms is not always necessary, since it can be bene cial for the RS materials to posses slightly di erent structure or defect concentration. Nevertheless, there is still a long way from the basic investigations and prototypes to the commercial applications. It is also important to be more careful measuring the electrical properties of the RS material, since the properties of the investigated material can change during the experiment itself, leading to unreliable results. Outlook Investigation presented in this work showed that the RS is complicated phenomena. The complexity of the RS even increases, when one moves his attention from the macroscale, typically used to describe the properties of the materials, into nanoscale. In many cases, only the nanoscale experiments can provide crucial data necessary for better understanding of the fundamental physics. What is more, the observations in nano-scale clearly shows that the perfect crystals are really not so perfect as one could expect, which is strongly connected to structure defects and inhomogeneity of the dopant. The knowledge of the imperfection and the real state of the materials is not only useful from the fundamental point of view. I strongly believe it is necessary for the future investigation and the development of new technologies. As the case of the STO (and the Fe doped STO) shows, the extended defects, such as dislocations, are crucial for the RS behavior. Therefore a careful investigations, mostly in nanoscale, can provides the ideas and the means to exploit the defects and imperfections for the bene t of new materials with outstanding properties.