This Materials World Network (MWN)/Inter-American Materials Collaboration project expands and deepens current understanding of correlated electron behavior in a variety of low-dimensional systems, including atoms and molecules on surfaces, electrons and spins in carbon nanotubes and graphene, and excitons in semiconductor quantum dots. Efforts focus on elucidating the roles of local and global symmetries, as well as interactions with thermal baths or radiation fields, in determining properties of interest for potential applications in nanomagnetism and spintronics, in new carbon-based devices, and in quantum information. A central goal is to guide and interpret experiments involving techniques that probe local interactions with high spatial and temporal resolution. This research is carried out by combining the complementary theoretical and numerical expertise of groups at Ohio University, the University of Florida, and Oakland University in the U.S., at the Pontificia Universidade Catolica do Rio de Janeiro and the Universidad Federal Fluminense in Brazil, at the Universidad Nacional de Rosario in Argentina, and at the Universidad de Antofagasta in Chile, with the non-US participants receiving support from their respective national funding agencies for their contribution. In addition to applying existing state-of-the-art methods, the team develops new techniques for the description of non-equilibrium phenomena.
The project immerses young scientists in cooperative international research activities through extended visits to MWN partner institutions and participation in a mini-workshop to be held in the second year of the award. Webcasts of technical lectures across institutions will advance technical training and foster collaborations among MWN participants. The project also seeks to engage a broader audience through the development of pedagogical materials related to nanoscience. Some of this development is carried out in collaboration with students from writing and visual design programs.
Two major outcomes have resulted from this project: i) development of a new way of doing calculations for systems of interest in nanotechnology ii) better understanding of specific systems related to spintronics. i) Miniaturization in electronics is forecast to reach the molecular level soon. This means that ways of creating and controlling devices involving a few atoms have to be developed. This can only be achieved if reliable ways of doing calculations at this level are developed. Efficient control of these devices is only possible if the models describing them are realistic yet simple enough that they can be handled by current computers. We have developed a way of exactly transforming a realistic model into an equivalent one that can be handled with great numerical accuracy. Consider Figure 1. The top panel describes a system of great interest in nanoscience: two magnetic atoms laying on the surface of a metal film. This can form the central part of a mesoscopic device. Understanding how the details of the surface influence the operation of the device is very important. The main idea is that the metallic surface mediates the interactions between the atoms, which is what makes the device operate. The bottom panel describes a representation of this system. This simplifying step is necessary, because this representation can be handled by a computer, the real system cannot. Note that the metallic surface was transformed into a chain of atoms, to simplify the calculations. The problem is that the majority of numerical tools that use this 'reduction' step only capture the gross features of the metallic surface, and more crucially, have to make assumptions about the interaction between the atoms. The method developed in this project performs this necessary transformation without losing the details. In this sense, it is called an exact mapping, and not just a 'representation'. More importantly, no assumptions are made about the interactions between the magnetic atoms. Whatever they are, they will not be modified. By varying the relative position of the atoms over the metallic surface, and subsequently probing the properties of the mapped system, one can understand, for example, how the interaction depends on the details of the positioning of the atoms over the surface. A few characteristic surfaces were probed through this mapping and their differences analyzed in the published results. ii) Current methods of storing and manipulating information inside a computer rely mostly on the flow of charge (electric currents). The heat dissipation generated by this paradigm is a severe constraint on further steps on miniaturization: the more transistors are packed per square millimeter, the more heat is produced per area. Electrons, beside charge, also carry spin, which can be thought of as a little magnet. Prevailing transistor technology does not harness the existence of this degree of freedom, which ideally can be manipulated with lower energy cost and therefore with less heat dissipation. Spin polarized currents will be a necessary ingredient of this new paradigm. Efficiently generating them is a very active field of research. Our project has suggested an efficient way of producing polarized currents. Consider Figure 2. In it, circles 1 and 2 schematically represent so-called quantum dots: man-made artificial atoms, whose number of electrons can be changed at will, one by one. By exploiting an effect discovered in the 1950s, the Kondo effect, involving magnetic atoms embedded in a metal, and which can be replicated and controlled using quantum dots, we show that it can be used to produce spin currents (charge currents where each spin points in the same direction) of opposing polarities (see right side of top panel) out of unpolarized currents (left side). This 'sorting out' of spins is accomplished by the specific properties of the Kondo effect involving both quantum dots in the central part of the device. In the process of studying how the Kondo 'state' achieves this sorting out, it became clear that there was a 'spatial splitting' of the state between the two dots: this can be seen in the bottom panels in Fig. 2, where we represent each dot by a quantum-well (where the electrons are stored) and show the calculated shape of the Kondo state. In the majority of Kondo effects this shape is the sum of the purple and red shaded curves, and any attempt to separate it (by applying a magnetic field, for instance) destroys the state, rather than splitting it. Here, by the very nature of the device and the associated Kondo effect, the state is naturally separated. In the published results, we detail how this polarization effect can be experimentally observed and also discuss a way of possibly verifying the spatial separation of the Kondo state.