Projects / Programmes
The many-impurity problem
| Code |
Science |
Field |
Subfield |
| 1.02.01 |
Natural sciences and mathematics |
Physics |
Physics of condesed matter |
| Code |
Science |
Field |
| P260 |
Natural sciences and mathematics |
Condensed matter: electronic structure, electrical, magnetic and optical properties, supraconductors, magnetic resonance, relaxation, spectroscopy |
| Code |
Science |
Field |
| 1.03 |
Natural Sciences |
Physical sciences |
electron correlations, dynamical mean field theory, magnetic adatoms, quantum impurity physics, magnetic data storage, numerical renormalization group
Researchers (8)
Organisations (2)
Abstract
Magnetic nanostructures that are relevant for future technological applications such as ultra-dense magnetic information storage are systems consisting of an intermediately large number of magnetic moments (100-1000) forming partially ordered adsorbate structures supported on solid substrates. In particular, the next generation in hard disk drives (estimated to be fully developed and commercialized toward 2025) is expected to be based on bit patterned media (BPM) where each bit is stored in a single-domain magnetic nanoparticle formed of the order of 1000 atoms. Feasibility of such devices has already been demonstrated by one major storage company, HGST. On the level of basic research, this calls for more thorough studies of the effects of magnetic anisotropy, inter-impurity exchange coupling, role of the substrate, partial positional order of adatoms, and superparamagnetism. Magnetic systems of this intermediate size pose theoretical challanges which sets them appart from problems involving a small number of impurities (such as single magnetic adsorbates, or cluster of two, three, up to four impurities) or a very large number of moments forming essentially a bulk (macroscopic) spin lattice system.
We will develop efficient, reliable and accurate numerical many-body techniques for studying systems consisting of a large number of quantum impurities (in the range of at least tens to hundreds, possibly up to thousands) in contact with a common bath of itinerant electrons which mediate effective inter-impurity interactions. These are models for large assemblies of magnetic adatoms on surfaces. We will focus on the most relevant impurity Hamiltonians, such as local spins with exchange coupling to the bath (as in the Kondo impurity model) or sets of orbitals with on-site electron repulsion and Hund’s coupling hybridized with the bath states (as in the Anderson impurity model), with extensions for incorporating additional effects that are particularly pronounced on surfaces, such as the magnetic anisotropy. Different types and phases of the bath will be considered (normal metal, doped Mott insulator, superconductor, topological insulator), since the nature of the dominant substrate-mediated magnetic interaction and its strength is determined by the low-energy two-particle excitations of the substrate. For such many-impurity systems we will compute local density of states (LDOS) at an arbitrary point in the system with sufficiently high spectral resolution to be compared with the dI/dV spectra measurable with low-temperature scanning tunnelling microscopes (STM), which are presently the dominant experimental tool to probe nanomagnetic islands on surfaces. We will also consider the potential energy landscape as a function of the collective magnetization of the whole structure and study the existence of possible long-lived metastable states.
The approach, based on the real-space dynamical mean-field theory (real-space DMFT) and its non-local extensions, will be applicable to arbitrary impurity systems and implemented in the form of an integrated software package. In the scope of the project, we will perform large-scale calculations for two paradigmatic cases:
1) the emergence of band-structure with heavy fermion excitations in Kondo lattices, starting from single Kondo impurities toward very large dense islands;
2) adatom clusters on doped Mott insulating (bad metal) and antiferromagnetic surfaces.
It is presently unclear what new physics one may expect in the range of hundreds of adsorbates (beyond shear complexity). Can we find unifying themes and unanticipated phenomena arising from coupling and competing interactions? The possibility of answering this question is a very exciting prospect.
Significance for science
The huge appetite for data storage capacity is a strong incentive for developing novel methods to increase the information density. The hard disk manufacturer HGST (a Western Digital company) has recently released a 8 TB drive to the market. To achieve such high capacity, they fill the drives with helium in order to reduce friction and vibrations. Another novelty is the shingled magnetic recording (SMR), pioneered by Seagate, where the data tracks partly overlap and read-update-write cycles are needed to store information. The next big step being considered is the bit-patterned media (BPM): instead of using conventional granular media, nanoimprinting is used to form a dense pattern of monodomain magnetic islands 10 nanometres across. HGST expects this technology, currently at the level of a prototype, to become cost-effective by 2025. This island size is near the upper limit of our target for the many-impurity solver, thus this project will, in fact, provide a methodology for studying magnetic structures that will most likely constitute the next milestone in the data storage industry for the decade after 2020, probably in the range of hundreds of atoms. At the end of this project, the research group will thus be in the position to actively contribute to the reseatch and development of the next stage of miniaturization that will start just around that time.
The shift from studying small nanostructures to exploring somewhat bigger assemblies is bound to become a trend also in the physics of strongly correlated systems. This project will constitute an early step in this direction, both by providing new tools and by setting direction through first applications to model systems. In long term, we expect that the tools will be developed to the point of being generally applicable and will be widely used to study other systems, also by other groups. This will be stimulated by open sourcing the tools, providing full examples (input files, scripts for running the calculation, as well as scripts for postprocessing), and organizing training sessions. In this area, the PI has had a nice experience with a training school for my NRG package organized in 2013 at SISSA, Trieste. The materials (presentations and the extensive examples) are publicly available and have turned out to be very useful to researchers that need to get started fast, as evidenced by the feedback received. We believe this is the best way to distribute scientific software, particularly in niche areas.
Significance for the country
The outcomes of this project would be useful for the sector of magnetic data storage. This industry includes companies with yearly income on the scale of 10 billion euros (Western Digital, Seagate Technology). These companies obviously have their own research & development departments, which are however also doing very basic research in the relevant fields. For instance, some of the most important studies of magnetic anisotropy in magnetic nanostructures on surfaces have been performed in IBM's Almaden research facility in USA (Andreas Heinrich group).
Most important scientific results
Interim report,
final report
Most important socioeconomically and culturally relevant results
Interim report,
final report
