IFF
Scientific Report 2000/2001

General Description of the Research Program

The research program of the IFF institute 'Electronic Properties' is devoted towards the investigation of the electronic structure of atoms, clusters, nanostructures, and solids. The ultimate goal is the development of an understanding and thus a base for the control of the properties of (new) materials. The electronic structure constitutes the microscopic base for all materials properties. The electronic interactions determine whether a solid is metallic, insulating, or semiconducting, whether it is transparent or exhibits a distinctive color, whether it is a magnet or a superconductor. Even the elastic properties and the thermal conductivity and heat capacity are determined by the electronic structure.

The electronic structure of simple solids consisting of a regular lattice formed by one or two atomic constituents, like pure metals or semiconductors such as GaAs, is quite well understood. Thin film systems introduce a modification of the intrinsic materials properties due to the interaction at the surface and the interfaces. In structures where at least one of the dimensions is in the nanometer range, the interface or surface properties in conjunction with quantum size effects result in significant modifications of the materials properties. This has led to the concept of 'atomic engineering of materials', whereby the materials properties are controlled by a variation of the atomic constituents at the microscopic atomic level. These concepts establish the central and interconnecting aspect of our research program, where we use both experimental as well as theoretical tools for our explorations. In our research we are not only interested in the ground state properties but also in the dynamic short time response to an external stimulation, as for example a short, intense light pulse.

A large part of our research program is devoted towards the development of a microscopic understanding of classes of materials with a direct connection to a technological application. Such materials where the microscopic control and the variation of the atomic constituents have a dominating influence for the technological application include semiconductor heterostructures as well as thin film magnetic systems for sensors and storage media.

Oriented towards these technological applications the research efforts of the institute are organized in the following areas:

• Magnetism of thin Films and Nanostructures; from Basic Research to Magnetoelectronics
• Clusters as new Materials

Additionally in our institute there is a considerable effort to develop new

• Methods and Instrumentation

Hereby the application of synchrotron radiation plays a central role. This includes the design and construction of beamlines and instruments for research with synchrotron radiation. In the following these programs are described in more detail. At the end of each section current research highlights are listed which either are included as short reports or have appeared in print recently.

Thin film systems and nanostructures such as magnetic quantum wires or dots exhibit distinctively different magnetic properties than their bulk counterparts. Obviously there is a close connection between the control of the magnetic properties of these novel nanostructured systems and the understanding derived from basic research on an microscopic atomic level. This has recently led to significant advances in technological applications of thin film systems as magnetic sensors and storage media. Additionally, this development is in line with the demand put forward by the increase in miniaturization in information technology.

Our long range goal is to develop a microscopic understanding of the magnetic phenomena. Magnetism is an electronically derived collective phenomenon influenced by the structure, composition and dimensionality of the system. As experimental techniques we are using spin and angle resolved photoemission, circular dichroism, and other synchrotron radiation related techniques in combination with (relativistic) band structure calculations. With spin- and angle resolved

photoemission we determine the bandstructure E(t,k) of the occupied electronic states. These data can be directly compared with bandstructure calculations based on the density functional theory. Element specific magnetic information about multilayer and multicomponent systems is obtained making use of the circular dichroism at core level absorption edges. In addition these studies yield element specific information about the collective ordering phenomena as a function of temperature.

As outlined above, electronic structure theory is an integral part of our research program and central to the development of a microscopic understanding of magnetic materials. The existence of a magnetically ordered phase is determined by a delicate balance between the exchange interaction and the kinetic energy of the electrons. The symmetry of the magnetic order (ferromagnetic or anti ferromagnetic) is also determined by the electronic structure and the geometry of the system. Spinorbit interactions, which in general contribute only a small amount to the total energy of the system, determine macroscopic parameters such as the magnetic anisotropy, the coercivity, and the magneto optical properties. The existence of a preferred orientation also leads to the breaking of symmetries and a lifting of the degeneracies of electronic states. At the present level most calculations do not yet incorporate these effects and efforts are undertaken to include some of these, for example noncollinear spin structures, into the calculations in order to get a more realistic treatment.

• Real Space Imaging of Two-dimensional Antiferromagnetism on the Atomic Scale
  S. Heinze, M. Bode, A. Kubetzka, O. Pietzsch, X. Nie, S. Bliigel, R. Wiesendanger, Science 288, 1805 (2000)
• One Dimensional Spin-Polarized Quantum Wire States in Au on Ni(110)
  C. Pampuch, O. Rader, T. Kachel, W. Gudat, C. Carbone, R. Klasges, G. Bihlmayer, S. Blügel, W. Eberhardt, Phys. Rev. Lett. 85, 2561 (2000)
• Electronic bandstructure of Gd: a consistent description
• Development of a non-erasable magnetic security strip and testing of prototypes

Magnetoelectronics is one of the key areas of information technology with a large potential for future expansion. In general the term magneto electronics has been coined to describe electronic devices where, in analogy to the microelectronics, the spin dependent charge transport is used for data storage and processing as well as for sensor applications. In the year 1999 we were successful in obtainig approval and funding for this project within the framework of the HGF-Strategiefonds.

The roots of magnetoelectronics can be found in the discovery of antiferromagnetic coupling in thin magnetic films separated by a non-magnetic metallic spacer and the "giant magneto resistive" properties of these multilayers by P. Grünberg (FZ Jülich) and A. Fert (Paris) in the late 1980's.

Even though it was not part of a technology roadmap or forecast, this unexpected quantum mechanical effect in the area of nanostructured devices has made its way from basic research to commercial devices in the amazingly short time span of less than 8 years. New types of magnetic field sensors were developed which are incorporated into the latest generation of readout heads for magnetic discs. This has resulted in an unexpected strong increase in storage density, since these new heads have a much higher sensitivity. New rotary sensors are being developed which are of interest in robotics and for multiple applications in automobiles. Further applications of the GMR sensors may be in non-destructive testing of structural materials, such as reinforced concrete, or of high speed train wheels, with high spatial resolution. Additionally prototypes for novel non-volatile magnetic random access memory devices (MRAM) are being developed, based upon the resistance change in magnetic tunnel junctions (TMR) upon reversal of the magnetization of one of the layers.

GMR sensors with applications as readout heads in hard discs or video recorders and as sensors in car-electronics guarantee a high volume and large revenues. The MRAM development on the basis of the TMR effect offers the possibility to build all solid state memory devices, which are nonvolatile and directly compatible with integrated CMOS technology. The worldwide market for (semiconductor) memory devices amounted to 38 Billion $ in 1998 and is projected to increase to more than 100 Billion $ in the year 2002.

The main goal of the HGF-program is to carry out a series of focused basic research projects in order to improve and expand the knowledge and materials basis for GMR sensors and TMR MRAM devices which are currently being developed jointly with industry within the framework of the "BMBF Leitprojekt Magnetoelektronik". The HGF-program is organized as a joint effort bridging institute boundaries whereby scientists from the IFF, the IGV and the ISI participate in this research. These novel magnetoelectronics devices are functional nanostructures consisting of ultrathin metal or metaloxide films, which have a thickness of only a few nanometers. Accordingly, quantization phenomena modify the electronic states and thus influence the magnetic properties. The alternating exchange coupling in magnetic thin films separated by a non-magnetic metallic spacer, which was at the origin of the development of GMR sensors, is indeed caused by the energetics of quantized conduction band states of the metallic spacer layer. Therefore it is obvious that the morphology of the layers and interfaces and their electronic properties determine the device characteristics. With respect to layout development questions arise concerning the transition into the superparamagnetic behavior with increasing miniaturization. Furthermore cross talk and the general influence of the lateral shape of the memory cells need to be investigated.

Until recently, the development of actual GMR sensors was largely empirical by trying out a large number of material combinations and process conditions. The development of TMR elements on the other hand is even less advanced. For both kinds of devices even the basic limitations for the optimum achievable performance, which are subject to the proper choice of materials and the modification of the interfaces, have not yet been established.

In the case of GMR sensors, for example, the relationship between the resistivity of the layers and the interface roughness has not yet been quantified. The development of artificial antiferromagnetic structures, such as FeMn multilayers is just at the beginning. Furthermore the influence of an antiferromagnetic insulating substrate, such as NiO, on the characteristics of the sensor is largely unexplored. Currently NiO has been used to pin the magnetization of one of the sensor layers and was found to also increase the GMR effect.

Concerning TMR junctions for the MRAM development, for the influence of the height of the tunneling barrier and even in general the variation of the spin dependent tunneling current upon magnetization reversal are not explained on a quantitative scale. Another largely unexplored, but nevertheless very important area, concerns the magnetization dynamics of thin film systems for MRAM cells. Furthermore, so far all TMR developments have used A1₂0₃ as the oxide barrier,

mostly due to historical reasons. Different oxides might lead to an improvement of the TMR characteristics.

Scientific highlights of the year 2000 (IFF only)

• Interface magneto-crystalline anisotropy probed by dynamical diffraction of soft X-rays
• Oszillatory interlayer exchenge coupling across an epitaxial FeSi spacer and ist unconventional temperature dependence
• Influence of the anisotropy on the FMR frequency of polycrystalline pinned and unpinned thin permalloy films

Our cluster research program exhibits quite some apologies to the corresponding research programs in solid state physics (magnetism, electronic properties) and surface physics (chemisorption, catalysis). This is in accordance with the main goals formulated above for our research program to develop an understanding for the properties of new materials on the microscopic atomic level.

Clusters of a monoatomic size are quite difficult to produce and separate in large quantities. Accordingly the investigation of the electronic properties of these individual mass selected clusters requires methods far beyond the capabilities of conventional electron spectroscopies.

Laser excited photoemission from anionic mass selected clusters has turned out to be a method which can be appplied to this task on a large variety of clusters in a mass selected molecular beam experiment. The clusters are produced by condensation of a plasma generated by a focussed laser beam or an electric discharge in a high pressure He atmosphere. Following adiabatic expansion and the formation of a molecular beam the cluster anions are accelerated by a pulsed electric field. According to their mass difference the anions thus arrive at different times in the ionization region of the magnetic bottle electron spectrometer. Adjusting the timing sequence of the laser used to excite the photoelectrons (detachment) electron spectroscopy on clusters of an unique mass can be performed. The electron energy distribution spectra reveal the electronic structure of the individual clusters. Under favorable conditions vibrational substructures may be resolved which reveal characteristic vibrational modes of the cluster and thus give additional information about the geometry of the atomic positions in the cluster. For the interpretation of these spectra and the identification of unique cluster structures a comparison with high level calculations and a collaboration with theory plays an important role.

In the year 2000 the major emphasis was shifted towards using femtosecond pump-probe techniques to study the mechanisms of energy transfer and relaxations processes. Here energy relaxation processes may be followed in real time for systems where the total energy remains localized spatially to a few atoms and correspondingly only a few degrees of freedom. Thus these processes evolve quite different than in solid state systems, where the total energy is spread rapidly over an extended volume of the sample involving more and more atoms. The energy can only be removed from the cluster by the emission of an electron or photon or by evaporation of an atom. This may occur finally after the system was evolving over a timescale up to the msee range. On the other hand very fast fs electronic relaxation processes may be observed also including the dissociation. Thus one is able to follow the development of the electronic structure as the fragments separate. This enables us to gain data about short lived transition states, which govern the pathways in chemical reactions. Accordingly these states are essential in developing a deeper understanding of chemical reactions. For the development of this field of fs-chemistry, A. Zewail was awarded the 1999 Nobel Prize in chemistry.

• Ultrafast hot-electron dynamics observed in Pt3' using time resolved photoelectron spectroscopy N. Pontius, P. S. Bechthold, M. Neeb, W. Eberhardt, Phys. Rev. Lett. 84, 1132 (2000)
• Pump-probe photoelectron spectroscopy on mass-selected metal cluster anions
• Time-resolved Photofragmentation of Metal Carbonylclusters

For technological applications the individual properties of the clusters have to be conserved. This can be achieved by depositing mass selected clusters onto a suitable substrate. For this purpose we have developed a high intensity source capable of delivering a sufficient amount of mass selected clusters onto a substrate within a few minutes to prevent contamination. Cluster ions are produced in a specially designed high intensity laser vaporization source. The cluster-ions present in the molecular beam following the adiabatic expansion are accelerated and deposited onto a substrate after mass selection by a sector magnet. The substrate can be prepared and characterized in a conventional XPS/UPS system before and after the deposition. Samples are exchanged between these systems using a small, portable UHV sample transfer chamber. This way the substrates can also be transferred to additional facilities, foremost STM and STS characterization in-house as well as spectroscopy with synchrotron radiation.

Typical areas of applications for deposited mass selected clusters are in catalysis or nanoelectronics. In catalysis it is well known that small particles may exhibit special properties, increasing the reactivity or more important the selectivity of the process. It is expected that these properties can be enhanced or tuned by mass selecting the catalyst particles. In electronic applications (nanoelectronics) the concept of a `single electron transistor' can be realized using deposited clusters. This is due to the fact that the clusters have discrete quantized electronic states, whereby the 'Fermi level' of a small cluster may change by an energy on the order of 1 eV, when a single electron is added to it or removed from it.

• Characterization of the properties of the deposited clusters --- including the atomic geometry and electronic structure in comparison with the `free' cluster in a molecular beam, diffusion and agglomeration phenomena, and generally the interaction with the support material
• Catalytic properties of mass selected deposited clusters
• Electronic properties of nanocontacts on semiconducting substrates
• Magnetism of nanostructures

• A 'metallic' endohedral fullerene: La@C_6o
• R. Klingeler, G. Kann, I. Wirth, S. Eisebitt, P.S. Bechthold, M. Neeb, W. Eberhardt (submitted to Nature)
• Scanning tunneling spectroscopy of size selected endohedral fullerenes

Both aspects, the design and realization of new instruments as well as the development of novel methods for spectroscopy, are part of our research efforts. The powerful infrastructure provided by a large research center is especially helpful when designing new and complex instruments. Our interests in this area are centered around the application of synchrotron radiation. On the hardware side this includes the design and construction of synchrotron radiation beamlines or specific spectrometers and detectors.

Additionally new methods for spectroscopy and scattering are also being explored, concerning the aspects of resonance excitation conditions as well as exploiting the coherence of synchrotron radiation. In the laboratory two-color pump-probe photoemission spectroscopy with fs timeresolution using lasers is a major experimental development. At the future FEL-sources the synchrotron beam properties will match the performance of laboratory laser sources with respect to coherence, pulse power, and timestructure (pulse-lengths) in the fs range. This will extend the capabilities we currently have at visible and UV photon energies into the VUV and (soft) X-ray range and open up the possibilities for a whole set of new and exciting experiments.

We have installed an undulator beamline for the photon energy range from 10 eV to 300 eV the DELTA storage ring. This beamline was planned for bandstructure investigations and high resolution core level spectroscopy of magnetic thin film systems for magnetoelectronics, solid state systems and cluster materials. We have built a plane grating type monochromator, since this type of instrument is, for the selected photon energy range, best suited to match the performance of the existing undulator source at DELTA. Unfortunately DELTA so far was unable to offer stable and reproducible operational conditions.

Additionally in the year 2000 we have installed an undulator based beamline with a spherical grating monochromator at the BESSY II storage ring for the photon energy range from 50 eV to 1.5 keV. The source consists of a pair of undulators offering adjustable polarization of the light, which is extremely useful for element specific investigations of magnetic systems including the determination of magnetic moments by using the sum rules developed for CMXD. We will also employ resonant inelastic X-ray scattering at this new BESSY beamline, a method which we have developed over the past years extensively. This beamline produced a first calibration spectrum in December and is currently being commissioned.

In the past year we have built up a new experimental station to conduct scattering experiments with coherent soft X-ray photons. This work was featured as the cover story in Synchrotron Radiation News. Even before the reconstruction of the interference pattern observed has been succesfully demonstrated last year, we had the vision that this might become a very interesting field of research in the future for the study of dynamical processes (m-sec) in magnetic systems, semiconductors and polymers with an exceptional sensitivity tunable on the length scale ranging from a few to 100 nanometers. As a special progress it should be mentioned that we recently were able to demonstrate that linear dichroism yields sufficient contrast to identify the axial orientation of domains on the surfaces of antiferromagnetic materials.

• Development of RIXS as a method to determine the bandstructure of solids
  (review article by S. Eisebitt and W. Eberhardt), J. Electr. Spectrosc. 110-111, 335 (2000)
• Speckle in coherent Resonant Soft X-Ray Scattering obtaining antiferromagnetic contrast
• Further advances in investigations of hydrophobic fluoro-alcyl-silanes

Time resolved pump probe photoemission spectroscopy is a new and very exciting field of physics. These experiments directly reveal the electron scattering and relaxation processes following the creation of an optical excited state in real-time. The shortest observed timescales of a few fs are due to electron electron scattering, whereas electron phonon scattering involves slightly larger timescales in the sub-ps regime. These processes also serve to populate electronic conduction band states not necessarily accessible by direct optical excitation. Recording the photoelectrons in an angle resolved mode allows the observation of the scattered excited electrons at specific points of the bandstructure. Since the excited electrons are observed directly, this technique does not require the presence or existence of recombination centers, which are exploited in ultrafast luminescence studies.

• Dynamics of image potential states on the Ni(111) surface

The theory activities of the institute are largely concentrated in two areas. One of them is the theory of magnetism in thin films including electron transport phenomena in magnetoelectronics as introduced above and the second topic concerns the theoretical modeling of STM images. This latter activity starts with the calculation of the electronic structure of the surface under investigation and the calculation of the tunneling current into a tip represented by an atomic s-, p- or d-wavefunction which is supposed to resemble the tip atom mostly responsible for the tunneling process. In the future a more accurate inclusion of the electronic structure of the tip is planned. However, even on the basis of this simple Tersoff-Hamarmn model (s-wavefunction) effects like the corrugation reversal of metallic substrates and the contributions to the STM image from atoms buried below the surface of metals could be observed, helping in the interpretation and correct assignment of some experimental puzzles. Previously these effects had only been reported for semiconductor surfaces, where electron screening is much less effective.

In the future we also plan to conduct experiments along these lines of the image development of the STM using our newly purchased variable temperature STM. Together with the theoretical activities we want to furnish compelling experimental evidence that even for metallic systems the STM images are substantially determined by the details of the electronic structure of the surface.

• Real Space Imaging of Two-dimensional Antiferromagnetism on the Atomic Scale

S. Heinze, M. Bode, A. Kubetzka, O. Pietzsch, X. Nie, S. Blugel, R. Wiesendanger,
Science 288, 1805 (2000)
• Resolving complex atomic scale spin structures by SP-STM

I hope that I could give a satisfactory overview of our research program and the common thread linking all these activities. Foremost in importance however are the people engaged in these activities and therefore I would like to conclude with a few general remarks about the changes among the members of the institute

Dr. C. Carbone has left the institute after more than a decade of highly interesting and successful research on the determination of the electronic and magnetic properties of thin magnetic films, nanostructures and multilayers. He has been appointed as Professor and group leader at the Institute for the Structure of Materials in Trieste. We wish him all the best for the future.

Dr. S. Blügel has an offer to be appointed as professor in theoretical physics at the University of Osnabriick.

I myself have an offer for the position as Scientific Director of Bessy in conjunction with a professorship at the TU Berlin.

Prof. M. Olmstead from the University of Seattle has received an A.von Humboldt award and has spent several months with us working on the growth and dynamics of semiconductor nanostructures.

Dr. N.V.Smith from the Advanced Lights Source in Berkeley, has also won this distinguished award from the A. von Humboldt society and is spending half a year with us collaborating in the area of magnetic nanostructures and fs spectroscopy of image potential states.

Dr. K. Maiti and Dr. S.S. Yan have each won a stipend from the A. von Humboldt foundation. Dr. Maiti works with us on the determination of the electronic structure of metal-oxide interfaces and Dr. Yan investigates magnetic thin film systems for magnetoelectronics.

Apart from these the following guests have spent extended periods time collaborating with us during the past year: Dr. G. Bihlmayer, Dr. S. Clarke, Dr. R. Gareev, Prof. B.Kuanr, Dr. S. Madsen, L. Sacharow, and Dr. X. Nie

A. Karl, P. Kurz, D. Olligs, and F. Voges have graduated and received a PhD during the last year and M. Breidbach, C. Friedrich, F. Förster, M. Lörgen D. Schondelmaier, and D. Wortmann have finished their Diploma thesis in the institute during the year 2000.

I would like to close this report by thanking all members of the institute for their efforts, contributions and discussions during the past year which have again resulted in a creative atmosphere and many scientific accomplishments

Wolfgang Eberhardt