Error! Style not defined.Error! Style not defined.

1

2

2.1

2.2Topic 2: Spin-based and quantum information

Spokespersons of the topicStefan Blügel

Topic costs2010: x M€

Personnelx scientists, y doctoral students,
z scientific support personnel

Contributing principal investigatorsG. Bihlmayer, T. Brückel, D.E. Bürgler, R. Calarco, T. Costi, P. Grünberg, H. Hardtdegen, R. Hertel, E. Koch, Ph. Mavropoulos, A. Liebsch, M. Lezaic, C. Meyer, E. Pavarini, U. Rücker, Th. Schäpers, C.M. Schneider, W. Schweika, K. Urban (NF), M. Wegewijs, D. Wortmann

This Topic explores phenomena and materials that may lead to devices that utilize the properties of an incoherent or coherent ensemble of electron spins or the static and dynamical properties of the strongly correlated many-electron spin as state variable or information carrier for storing or processing information. An example of the latter is the magnetization as order parameter of a ferromagnet, a system resulting from the collective behaviour of electrons with aligned spins. The Topic is part of the wider field of spintronics. As guiding principle of our research, we explore phenomena, investigate and pattern materials, and develop concepts and methods that

  • offer the preparation and detection of spin-currents and coherent spin states
  • allow for long spin-life time and slow decoherence
  • offer effective ways of controlling and manipulating spin- or magnetic properties by external parameters such as electric currents or gate voltages
  • switch the magnetization at the speed of the ultimate physical limit, local in space and with minimum energy impact
  • cross-correlate the magnetization to order parameters of other phases such as orbital ordering or the ferroelectric polarization aiming at devices with higher functionality or higher phase response

In order to come closer to the ultimate goal to make practical device schemes that will enhance the functionalities of current charge based systems or even to develop functionalities unattainable by current charge based systems.

Exploiting the full potential and extensive opportunities provided by the spin degree of freedom of charge carriers is a major goal in information technology. The ITRS road map clearly identifies spintronics as an emerging technology for the beyond-CMOS era, which holds the promise of new functionalities, higher speed and less energy consumption. First generation devices, such as sensors based on giant and tunnelling magnetoresistance have already revolutionized the data storage sector, pushing the storage capacity of hard disks well beyond the 100 Gbit/inch2 limit. Pioneering discoveries in this area have been honoured by the 2007 Nobel Prize in Physics for Peter Grünberg (Jülich) and Albert Fert (Paris).

2.2.1Challenges

Spintronics is a field enjoying a rich spectrum of fascinating fundamental discoveries and practical innovations. While discoveries like those related to the spin-Hall effect, to name just one, need to be understood in more depth and detail, and their consequences for future practical devices may not yet be clear, the growth of tunnelling junctions with a giant tunnelling magnetoresistance ratio define clear roadmaps of future innovation. We incorporate these new phenomena and develop spintronics along various thrusts relevant for different time horizons. First, an improved control of spin currents – and thus of spin-transfer processes in general – is required to bring about the next generation of devices with added functionalities, for example, magnetologic circuits, transistors, or microwave generators. Second, the replacement of conventional magnetic tunneljunctions by junctions with functional oxides and oxide heterostructures offers a potential for more functionality. Both thrusts imply the solution of many conceptual and material-related problems. Third, the long-term goal must be to develop means of manipulating individual spins and tuning and controlling their interactions by stretching spin lifetime and spin coherence in a condensed matter environment. This will be both a precondition and major step on the way to a future solid state-based quantum-information processing approach.

Representing the very forefront of nanoelectronics, spintronics rests on a firm basis of fundamental research activities, which may be divided into four main areas. These activities are very closely interrelated and at the same time define the challenges in the field:

(a) Nanomagnetism
Essential building blocks in spintronics are nanoscale magnetic elements, ranging from lithographically defined nanopatterns down to individual magnetic clusters and molecules at the spatial limit. Nanomagnetism stands for the development of strategies and procedures to design and manipulate the magnetic ground state, spin structure, and magnetic coupling phenomena in and between nanosized magnetic structures and objects. This includes joint research on new functional materials, e.g. magnetic molecules, withTopic 4.

(b) Spin transport and coherence
The main functionality in spintronics is provided by the manipulation of spin currents. The areas of spin transport and coherence deal with the unique issues of the preparation of spin currents, the transfer and charge transport phenomena, such as spin accumulation, spin injection, spin relaxation, spin torque transfer, and spin dephasing. Mastering these issues holds the main key to technological exploitation.

(c)Magneto- and spin dynamics
The temporal limit of reading and writing information in a condensed matter environment is not yet known. The functionality of a spintronics device depends crucially on the ability to switch the magnetization or manipulate a spin ensemble on short time scales. Therefore, by magneto- and spin dynamics smart we address switching strategies by spin-polarized currents, pure spin currents, spin waves or photon pulses and exploring the physical limits of the various dynamic regimes;

(d)Nanoferronics
Aiming at a route to multifunctional tunnelling contacts, allowing for example for a high electric field response of spin-dependent transport properties, passive oxide, e.g. MgO, barriers in conventional magnetic tunnelling elements are replaced by hysteretic oxides or oxide heterostructures, which display a rich variety of physical phenomena including magnetism, piezoelectricity, ferroelectricity, multiferroicity and superconductivity.

All phenomena in spintronics have a common microscopic basis, which results from electronic interactions and correlations. Understanding the ground state electronic structure and their excitations, particularly in complex systems, and even predicting specific electronic properties are the challenges to be mastered on the way to developing powerful design rules for spintronic systems.

The general research strategy and approach in this Topic is characterized by a particularly close interaction between theory and experiment, exploiting a broad portfolio of techniques, which are also continuously developed. These encompass on the one hand a full suite of spectroscopy, microscopy and scattering approaches, and on the other hand powerful theoretical methods ranging from electron theory, i.e. density functional theory and dynamical mean field treatments for the description of ground and excited states of condensed matter systems, to the modelling of complex micromagnetic structures by means of state-of-the-art finite-element simulations to study the static and dynamic properties of the magnetization between the nanoscale and mesoscopic length scales.

2.2.2Subtopic: Nanomagnetism
Current activities and previous work

The nanomagnetic systems studied range from individual atoms and magnetic molecules to thin film systems. We are interested in the magnetic properties of the individual magnetic entities and in particular in the interactions between them. Magnetic interactions span very different length scales, from the short-ranged exchange coupling to the long-ranged dipolar interaction. As a consequence, a wide variety of magnetic configurations and spin structures are formed, whose analysis is a problem far from trivial.

Figure 2-1:Mn monolayer film on W(110) with homochiral spin order due to Dzyaloshinskii-Moriya interaction. Mirror image below does not exist [2].

At surfacesThe Rashba effect is well known to occurin semiconductor-heterostructures and is caused by the spin-orbit interaction in structurally asymmetric environments. We found that it can also be observed on surfaces and thin films of heavy elements like the semi-metal bismuth [1]. Thus, even on nonmagnetic surfaces spin-dependent phenomena can arise, which are of considerable interest in the field of spintronics: In the potential gradient near the surface, electrons with different spin orientation have different band structures, making these structures potential candidates for spin-manipulators and possibly a first realization of a topological insulator besides the known band- and Mott-insulators. Electrons subjected to the Rashba effect cause an antisymmetric exchange interaction, named after Dzyaloshinskii and Moriya, of so far unknown strength. Employing relativistic noncollinear first-principles calculations we showed that the interaction is sufficiently strong to destabilize ferro- or antiferromagnetic order. We discovered a homochiral, i.e. single-handed, magnetic order in a two-dimensional Mn film on W(110) shown in Fig. 2-1 [2]. This is a rather general phenomenon, which has been overlooked the past 20 years.

Figure 2-2: Experimental (left) and simulated (right) reflectivity map of a Fe/Si multilayer with twisted magnetization state (middle) [5].

MultilayersExchange-coupled multilayers and heteromagnetic interfaces between ferro- and antiferromagnets give rise to competing interactions and thus unusual magnetic structures. Our laterally resolved x-ray magnetic dichroism study of the single-crystalline model system Fe3O4/NiO revealed pronounced proximity effects in both the ferro- and antiferromagnet and a strong dependence of the coupling character on the crystalline orientation [3]. Some exchange-bias systems reveal an asymmetric remagnetisation behaviour, where remagnetisation in one branch of the hysteresis loop is due to coherent rotation of the magnetisation, while in the other branch domain nucleation and growth occurs. Employing neutron reflectometry with polarisation analysis in an external field applied under different angles with respect to the unidirectional anisotropy of the exchange bias system allowed us to solve this puzzle and explain it with micromagnetic models and simulations [4]. In an antiferromagnetically coupled Fe/Si multilayer we have observed a twisted magnetisation state (Fig. 2-2) [5] and determined the field dependent magnetic domain structure in GMR multilayers.

Contents and goals

The investigation of spin-orbit related phenomena of atomic-scale films and nanostructures is a major focus for the coming years. The discovery of the Rashba effect and the Dzyaloshinskii-Moriya interaction at metal surfaces has opened many completely new vistas: (i) Since space inversion symmetry is broken at the surface, the spin degeneracy of surface bands can be lifted by the spin-orbit interaction. Under certain conditions this can lead to spin-polarized electron bands that acquire a non-trivial Berry phase and spin-textured edge Fermi contours that reveal a new quantum order, described by a mirror Chern topological number. This describes uniquely the topological insulators, which are a fundamentally new quantum phase of matter that may provide a route to fault-tolerant quantum computing. (ii) The investigation of the spin-orbit dependent scattering at atoms, molecules, self-assembled or man-made nanostructures enables the study and understanding of microscopic scattering processes that lead to spin-relaxation, dephasing and decoherence processes or act as extrinsic or intrinsic mechanisms responsible for the anomalous or spin-Hall effect. We explore the hypothesis, that some of these processes are fundamentally different in nano-scale structures as compared to the in bulk. (iii) The homochiral magnetic structures caused by the Dzyaloshinskii-Moriya interaction have a very low pinning energy and may be moved by the torque exerted by a spin-polarized current. Furthermore we search for Skyrmions – two-dimensional homochiral magnetic structures -- and Skyrmion lattices by exploiting certain symmetry conditions, e.g. considering a (111) oriented magnetic cluster on a substrate. (iv) Dynamic spin processes at surfaces are investigated by inelastic spin-polarized scanning tunnelling microscopy, the theory of which will be developed.

Miniaturisation of spintronic devices leads to the use of magnetic nanoparticle, where a transition occurs from ferromagnetic to superparamagnetic behaviour as a function of the particle size. Alternatively, magnetic molecules, whichstand out by the fact that due to the chemical synthesis road all particles are identical with the same number of spins have the potential for applications as the smallest practical units for magnetic memories or possible realisation of quantum bits (Qubits) in future quantum computers. It is a true challenge to determine the magnetic structure within such nanoparticle or the magnetization distribution within a magnetic molecule. These problems can be resolved by scattering methods since these do not disturb the magnetic state of the single particle. Small angle neutron scattering with polarisation analysis on nanoparticle systems with extremely narrow size distribution will allow us to determine the average structure and magnetic structure within such nanoparticle. An alternative approach is to use Pair Distribution Function (PDF) analysis employing both neutron and high-energy synchrotron radiation. Polarized neutron scattering will allow us to determine the magnetization distribution in molecular magnets and together with inelastic neutron scattering to determine the interaction and magnetisation dynamics in such nanoparticle systems. For spintronic devices, nanoparticle or molecular magnets have to be arranged in a periodic manner on a substrate. Scattering under grazing incidence enables the study of magnetic interactions and remagnetisation behaviour of such ordered nanoparticle arrays.

2.2.3Subtopic: Spin transport and coherence

The investigation of spin-dependent transport phenomena was pursued along three parallel routes, which are connected to the understanding and use of different material classes. The first route was concerned with the preparation of optimal leads for tunnelling magneto-resistance elements, in which the classical magnets are replaced by half-metallic ferromagnets, in particular, Heusler phases. The second route addresses in important issue in semiconductor spintronics, which is the manipulation of spin currents by electric fields. The third route prepares carbon nanotubes as quantum dots, a research activity started by the Young Investigator Group of C. Meyer. In the following research period we focus our effort to the realization of quantum dots and the understanding of spin-scattering, spin-decoherence and spin-dephasing mechanisms and processes.

Current activities and previous work
Figure 2-3: Weak antilocalization in GaInAs/InP quantum wires. The conductance maxima at zero field indicate clear weak antilocalization for wide wires. / Figure 2-4: SEM micrograph of a carbon nanotube filled with fullerenes, a so-called peapod. The inset indicates partial filling from top of the tube.

Metal-based spintronicsMagnetic tunnel junctions with Co-based Heusler alloys and MgO barriers reveal significant tunnelling magnetoresistance (TMR) at high bias voltages [6] as well as large inverse TMR effects, both contrasting the properties of classical TMR junctions and indicating the influence of the particular electronic structure of Heusler alloys on the tunnelling behaviour.

Semiconductor spintronicsBy studying the weak antilocalization effect in GaInAs/InP quantum wires we could demonstrate [7] that spin precession due to Rashba spin-orbit coupling is still present in strongly confined systems. For very narrow wires a suppression of weak antilocalization was observed, which could be theoretically modelled [8]. Triggered by the growing interest in diluted magnetic semiconductors based on GaN, we investigated the spin transport in AlGaN/GaN heterostructures. Weak antilocalization measurements prove that despite the large band gap spin-orbit coupling is still present.

Carbon nanotube quantum dotWith respect to coherent spin manipulation on quantum dots it is known for GaAs that photon-assisted tunnelling interferes with the desired spin resonance effect. As a first step towards local manipulation of spins in carbon nanotube quantum dots, we studied photon-assisted tunnelling in these structures [9]. In order to fabricate devices of functionalized carbon nanotubes for transport measurements, we investigated oxidation and filling (Fig. 2-4) of individual tubes on a surface.

Contents and goals

The separation of spin and charge currents in non-local spin-valve structures and the direct as well as the inverse spin Hall effect are important building blocks for the realization of a charge-less spin logic and its interfacing to the conventional charge-based electronics. In particular the spin-Hall effect and the quantum spin Hall effect are believed to add to future electronics a new dimension by exploiting spin currents flowing in the system without dissipation. At the current state of development of the field it seems necessary that in close connection to experiments extensive theoretical investigations are conducted with the focus on phenomena related to various types of Hall effects, and, generally speaking, phenomena in which the interplay of charge and spin degrees of freedom of electrons in nanostructured materials stemming from the so-called geometrical, or Berry phases, can be employed for future nanoelectronics applications. We plan developing this comparatively new concept in condensed matter physics and will provide valuable insights into spin-orbit driven phenomena of novel Hall type family transport properties on qunatitative basis. Lateral nanostructures with 4-terminals enable non-local transport measurements for the investigation of advanced spin transport concepts will be prepared.

Concepts for quantum information hardware are currently driven by theory. In order to explore the practical limits of quantum computers, it is necessary to understand the decoherence processes and mechanisms resulting from an interaction of the quantum hardware with the environment. Once the processes are understood to a certain extent,they can be simulated by formalisms implemented on massively parallel supercomputers (Collaboration with the Helmholtz Research Programme Supercomputing).

We investigate transport, interactions as well as decoherence channels of electron spins in systems with low dimensions, particularly one-dimensional systems, with respect to quantum information processing. For this, we study quantum transport on single electron quantum-dot structures based on semiconductor nanowires or carbon nanotubes (CNTs) including coherent control of the spin-state by locally employing microwave excitations. In nanowires spin-orbit coupling due to bulk inversion asymmetry or structural inversion asymmetry (Rashba effect) is of special interest. Results obtained from measurement in nanowires produced from 2-dimensional electron gases are used to find new ways to use the Rashba effect for spintronics combined with coherent control of the electron spins in epitaxially grown nanowires with core-shell structures. In CNTs the spin-orbit coupling is small compared to nanowires. However, chemical functionalization of the interior and exterior of CNTs allows one to tune their properties. In particular they can be filled with endohedral fullerenes, C60 with an atom or another small molecule inside, which carry an electron spin and thus, form a model system for studying one-dimensional spin chains. In order to characterize the functionalization of a nanotube device, Raman spectroscopy and aberration corrected TEM measurements (done with ER-C) on individual nanotubes are correlated. All these experimental studies are supported from first principles which provide microscopic parameters used to estimate the host-dependent amplitudes entering the decoherence rates and are combined to model theory such as dynamical non-equilibrium effects aiming at the quantitative understanding of (magneto-) transport and decoherence rates of electrons in weakly disordered wires containing magnetic impurities, such as Mn, Fe, or Co (see also Topic 4, challenge (b)).