Table of contents

Recent results obtained in our laboratories on interlayer exchange coupling of Fe films across interlayers of iron silicides, Fe_1−xSi_x with x = 0.5– 1, are reviewed. Samples are prepared by molecular beam epitaxy and characterized by means of low-energy electron diffraction and cross-sectional transmission electron microscopy. Coupling across interlayers of iron silicide with x ≈ 0.5 is found to be oscillatory with a strength of the order of 1 mJ m⁻², and across well ordered Si interlayers (nominally x = 1) the coupling is exponentially decaying. In the latter case the maximum coupling turns out to be surprisingly strong (> 6 mJ m⁻²), in particular considering the fact that the electrical resistivity is found to be large. Current–voltage curves for currents across the interlayers are characteristic of electron tunnelling. Soft-x-ray emission and near-edge x-ray absorption spectroscopy further support a semiconducting nature for the nominally pure Si interlayers.

The spectroscopic splitting factor g and the Gilbert damping constant G are magnetic parameters accessible to ferromagnetic resonance (FMR) measurements, which apart from the magneto-crystalline anisotropy energy can provide information on the spin–orbit coupling in magnetically ordered material. Whereas the effect of spin–orbit coupling has been thoroughly investigated and is well understood in insulating transition metal compounds, in 3d-metallic magnetic compounds the microscopic mechanism still needs further clarification. Particularly in thin films and multilayers interface effects and interaction between layers can modify both spin and orbital moments leading to changes of the g-value and the Gilbert damping constant. Experimental results are presented from frequency dependent FMR measurements on Co epitaxial films grown on Cr(001) and on films of the alloy Co_1−xFe_x(100) deposited on MgO(001), and from recent studies on Fe(100) films grown on InAs(001). The experimental data yield clear evidence of the importance of surfaces or interfaces of the films on the magnitude of orbital and spin moment.

Ferromagnetic resonance (FMR) is known to be one of the most informative techniques to measure basic physical quantities such as magnetic anisotropy energies, the g tensor in solids or the interlayer exchange coupling J_inter. We investigate prototype Cu/Ni/Cu/Ni/Cu(001) and Ni/Cu/Co/Cu(001) trilayers as well as Fe_n/V_m superlattices. We show for the case of trilayers how in situ ultrahigh vacuum FMR can be used to determine J_inter in absolute energy units in a straightforward way: we first prepare and measure the bottom magnetic layer together with the Cu spacer in situ and then evaporate the second magnetic film on top. Thus, it is possible to investigate the FMR signal before and after the two magnetic films become coupled. We discuss results, showing that the temperature dependence of J_inter follows a T^3/2 law over a wide temperature range. This indicates that thermally excited spin waves at the interface of the ferromagnetic layers dominate the temperature dependence of J_inter. The second part focuses on the measurement of the g value. From the g value, the ratio of orbital to spin magnetic moment can be obtained via the relation μ_L /μ_S = (g − 2)/2. We show for Fe_n/V_m superlattices how μ_L /μ_S increases with decreasing Fe-layer thickness.

Starting from an algebraic description of collinearity within density functional theory the concept of magnetic configurations is introduced, which in turn is of crucial importance to define the interlayer exchange energy (IEC) and the (giant) magnetoresistance (GMR). The applications shown are meant to clarify not only the conceptual basis of IEC and GMR, but also the difficulties that can arise in actual calculations.

We present theoretical studies of the spin wave excitations in ultrathin Fe films adsorbed on the W(110) surface, within the framework of itinerant electron theory, and with use of a realistic electronic structure. The electronic structure of the films and the substrate, taken as semi-infinite, is described by the empirical tight-binding scheme with ferromagnetism in the film driven by intra-atomic Coulomb interactions within the d shell. We calculate the exchange stiffness directly, then explore spin wave dispersion relations and Landau damping throughout the surface Brillouin zone through use of the random phase approximation. We also compare dispersion relations so generated with results based on the adiabatic description of spin motions; the latter approach overlooks Landau damping, of course, which we find appreciable at the shorter wavelengths. We comment also on apparent hybridization gaps and 'optical' spin waves which have appeared in earlier theoretical studies of spin waves in the bulk transition metal ferromagnets.

The current interest in the magnetism of ultrathin films and multilayers is driven by their manifold applications in the magneto-and spin-electronic areas, for instance as magnetic field sensors or as information storage devices. In this regard, there is a large interest in exploring spin structures and spin disorder at the interface of magnetic heterostructures, to investigate magnetic domains in thin films and superlattices, and to understand remagnetization processes of various laterally shaped magnetic nanostructures. Traditionally neutron scattering has played a dominant role in the determination of spin structures, phase transitions and magnetic excitations in bulk materials. Today, its potential for the investigation of thin magnetic films has to be redefined. Polarized neutron reflectivity (PNR) at small wavevectors can provide precise information on the magnetic field distribution parallel to the film plane and on layer resolved magnetization vectors. In addition, PNR is not only sensitive to structural interface roughness but also to the magnetic roughness. Furthermore, magnetic hysteresis measurements from polarized small angle Bragg reflections allows us to filter out correlation effects during magnetization reversals of magnetic stripes and islands. An overview is provided on most recent PNR investigations of magnetic heterostructures.

In general, elementary excitations in solids such as crystal field (CF) transitions and phonons are considered decoupled and the determination and interpretation of the measured spectra of the two phenomena, i.e. the CF level schemes and the phonon dispersion relations, are performed independently of each other. In addition, the spectra of these two excitations are generally quite complex and hence any unusual features are difficult to detect. A signature of a strong coupling between the two phenomena is the observation of an unusual behaviour in both subsystems. To prove the coupling unambiguously it is therefore necessary to investigate e.g. the phonon dispersion relations of an isostructural compound where the magnetic rare-earth ion (Ce, Yb) is replaced by a non-magnetic, but chemically equivalent ion (Y, La, Lu) and to determine the CF schemes of the same compound with the rare-earth ion replaced by a 'normal' magnetic rare earth. This requires, of course, time-consuming, detailed investigations. With these considerations in mind, it is not a surprise that there are only a few examples known where a coupling between electronic (CF transitions) and lattice (phonons) degrees of freedom have been reported.

Here we will discuss results on three rare-earth compounds where the coupling between CF transitions and phonons has been unambiguously shown by inelastic neutron scattering experiments: CeAl₂, YbPO₄ and CeCu₂.

Spin reorientation transitions (SRTs) of ultrathin Co/Pd(111) films induced by adsorption of atoms and molecules have been investigated by means of Co L_III,II-edge x-ray magnetic circular dichroism (XMCD). We have examined CO, NO, O and H as chemisorbed species. We observed the adsorbate-induced SRT from surface parallel to perpendicular magnetization at 200 K in the cases of CO and NO on 4.5 ML Co/Pd(111), while O or H adsorption has been found not to lead to the transition on 4.5 ML Co. We have investigated the detailed Co-thickness dependence in the case of CO and have found that the critical thickness of the SRT in CO-adsorbed Co is ∼6.5 ML, which is greater by ∼3 ML than that of clean Co (∼3.5 ML). From the sum rule analysis, it has been found that the surface parallel orbital moment is reduced after CO adsorption, while the surface normal orbital moment is left unchanged. We can conclude that the observed stabilization of perpendicular magnetic anisotropy due to adsorption is ascribed to quenching of the surface parallel magnetic orbital moment.

We report on the induced magnetism at the interfaces of layered structures of 3d and 5d metals explored with the x-ray magnetic circular dichroism (XMCD) technique. The beginning as well as the end of the 5d series were analysed by investigating W/Fe and Ir/Fe superlattices. W is found to couple antiferromagnetically and Ir ferromagnetically to the Fe layers. The induced spin moment is oriented parallel to the induced orbital moment for the W as well as the Ir case. For the early 3d element vanadium, a thickness-dependent study was carried out for prototype Fe/V/Fe(110) trilayers to study the range of the polarization in vanadium. A short-range polarization was found, resolving inconsistencies between earlier theoretical and experimental works. Since the sum rule analysis usually used for the XMCD technique is not appropriate for the early 3d element vanadium, the experimental absorption fine structure and the magnetic dichroism are compared to ab initio calculations in order to obtain magnetic ground state properties.

Figure 7 in this article was corrected on 21/03/2003 because of an error introduced during publication.

A depth-resolved technique is applied in the x-ray magnetic circular dichroism (XMCD) method by controlling the probing depth of the electron yield XMCD spectra. The usefulness of this technique is demonstrated for the study of magnetic structures of 4 and 8 ML Fe films grown on Cu(100), which are known to exhibit peculiar magnetic depth profiles. It was directly shown that the 4 ML film is uniformly magnetized, while the magnetic moment is localized at the surface in the case of the 8 ML film. The XMCD spectrum for each layer of the 4 ML film was separately extracted. All the extracted spectra were almost identical to each other, confirming the ferromagnetic coupling over the whole film. As for the 8 ML film, it was suggested that the surface two layers are ferromagnetically coupled, while the inner layers are in a spin density wave state with a wavenumber q = 2π/2.4d.

We have performed x-ray magnetic circular dichroism, x-ray resonant magnetic scattering and scanning tunnelling microscopy measurements on ultrathin Co films deposited on flat and vicinal Cu(111). The range of film thickness varies between one and 25 monolayers. For the vicinal Cu(111), Co deposition below one monolayer leads to the formation of elongated islands preferentially oriented along the step edges. These islands extend over lateral length scales of up to several thousand ångströms. For the thicker films we still observe that the vicinal surface leads to the occurrence of a preferential uniaxial growth direction. No such preferential growth direction can be identified for the flat surface. In correlation to the Co growth we observe an increase of both the orbital and the spin moment per Co atom on vicinal Cu(111) of about 25% versus what was observed for Co on flat Cu(111). This enhancement is observed in both the x-ray absorption and the reflectivity measurements. Similar to what was earlier reported for Co on Cu(100) we also observe for the case Co on Cu(111) an increase in the ratio m_l /m_s (orbital to spin moments) of 40% for thin Co films. In contrast to what has been reported for Co films on flat and vicinal Cu(100) we do not observe any major variations in the occupancy of the Co 3d states for Co grown on the vicinal Cu(111) surface.

First-principles density functional theory calculations have achieved great success in the exciting field of low-dimension magnetism, in explaining new phenomena observed in experiments as well as in predicting novel properties and materials. As known, spin–orbit coupling (SOC) plays an extremely important role in various magnetic properties such as magnetic anisotropy, magnetostriction, magneto-optical effects and spin-dynamics. Using the full potential linearized augmented plane wave approach, we have carried out extensive investigations for the effects of SOC in various materials. Results of selected examples, such as structure and magnetic properties of Ni/Cu(001), magnetism and magnetic anisotropy in magnetic Co/Cu(001) thin films, wires and clusters, magnetostriction in FeGa alloys and magneto-optical effects in Fe/Cr superlattices, are discussed.

We discuss in this paper the magnetic and structural parameters of Fe/V and Fe/Co multilayers. The electronic structure, magnetic moments (spin and orbital) and Curie temperatures as well as the magneto-crystalline anisotropy are calculated using first principles theory. Although theory is fairly successful in reproducing the experimental data we argue that the observed difference between theory and experiment most likely is due to lattice imperfections and that the interface between e.g. Fe and V is not perfectly sharp. We also present a model, based on the theory of elasticity, for analysing the structural properties of multilayers.

The response of a normally non-magnetic material to an external static magnetic field can be monitored by many different magnetic response functions. A general scheme to supply a corresponding theoretical description is presented that is based on the relativistic Korringa–Kohn–Rostoker Green function method of band structure calculations, and for that reason has many appealing features. First of all, it treats all spin as well as orbital contributions to the induced magnetization on the same footing, accounting in particular for all relativistic influences. Furthermore, it is extremely flexible and applicable in principle to any kind of system. Finally, a formulation for any response function can be worked out on this platform in a rather straightforward way. This is demonstrated together with corresponding applications for the magnetic form factor, the magnetic susceptibility and the Knight shift. In addition, a description of the so-called field-induced magnetic circular dichroism in x-ray absorption will be presented. In particular, it is shown that this new magneto-optical effect supplies a rather unique probe, giving information separately on the spin and orbital susceptibility in an element-resolved way.

In this talk I will give my personal view of important areas in magnetism research where x-rays have the ability to make key contributions. The selected areas deal with the manipulation of magnetization on the nanoscale with unprecedented speeds, i.e. on the picosecond and even femtosecond timescale. This development is driven by the need of information technology to store information more efficiently and process it faster. Implementation of this goal requires concepts that deviate from the classical concepts of switching the magnetization with Oersted fields created by current flow through wires or coils. The new concepts are based on the generation of spin-polarized electrons and spin states with a finite lifetime. For example, the generation of entangled spin states holds promise for futuristic concepts such as quantum computing, while spin-polarized currents can be used to switch the magnetization in nanostructures through the strong short-range quantum mechanical exchange interaction. I will demonstrate how such phenomena can be studied by means of advanced x-ray techniques that provide high spatial resolution, through real and reciprocal space imaging, in conjunction with ultra-short pulses. The envisioned x-ray studies utilize existing synchrotron radiation sources and, in the future, soft x-ray free-electron lasers.

Optical activity (OA) was only recently discovered in the x-ray range. It is caused mainly by the electric dipole–electric quadrupole E1E2 interference terms which mix multipole moments of opposite parity and exist only in structures with odd space parity. OA related phenomena can be either even or odd with respect to time reversal: natural OA is concerned with time-reversal even effects whereas non-reciprocal OA refers to time-reversal odd contributions. Various types of x-ray dichroisms related to either natural or non-reciprocal OA have been detected and are reviewed in the present paper. Extra emphasis is put on two different non-reciprocal effects that were observed in selected magnetoelectric materials: (i) x-ray magnetochiral dichroism (XMχ D) and (ii) the non-reciprocal x-ray magnetic linear dichroism. Edge-selective OA sum rules have recently been established which show that the effective operator of XMχ D is the orbital anapole moment at the absorbing centre.

There has been dramatic progress in recent years both in efficient calculations and in the interpretation of various x-ray spectroscopies. For example, real-space multiple-scattering theory gives a unified treatment of extended x-ray absorption fine structure, near edge structure, x-ray magnetic circular dichroism and other spectroscopies. Here we discuss the theory of x-ray spectra in the framework of excited state electronic structure. These are closely connected since x-ray spectra are directly related to a Green function for the excited photoelectron in the presence of a core hole. However, corrections to the independent electron approximation are needed to account for many-body effects such as inelastic losses, screening and core-hole effects. We discuss recent approaches for calculating such corrections and show that they can explain some of the remaining discrepancies between theory and experiment.

It is well known that there are close correlations between structural and magnetic properties of materials and certainly this holds also for nanostructured epitaxial films. So, the knowledge of the crystallography of a certain structure - i.e. the knowledge of the precise coordinates of the atoms involved - is essential for a quantitative understanding of a system's magnetic properties. Unfortunately, real space methods such as scanning tunneling microscopy provide crystallographic data of only rather limited accuracy and - even worse - only of the top layer. The full structure of a film can only be resolved by techniques applying surface penetrating probes, for example, x-rays or electrons (or both).

The present talk illuminates the power and limitations of using electrons or, more precisely, low-energy electron diffraction (LEED) in its quantitative version (applying tensor LEED for the intensity analysis). The talk concentrates on metallic epitaxial films of nickel, cobalt and iron on low index copper surfaces as well as iron on a reconstructed iridium surface. It will be shown that in favourable, i.e. structurally simple cases, quantitative LEED can resolve atomic positions with an accuracy of the order of 0.01Å, as well as the chemical nature of the atom under consideration. However, the accuracy reduces with increasing structural complexity. This is because of both correlations between parameters and, quite often, some lack of the scientist's imagination in considering all relevant parameters. Complexity can hold even in simple cases when different structural domains exist, with each of them to be analysed, but only the sum of the intensities is available for the fit. On the other hand, the method can be extremely sensitive to a parameter (for example, vertical layer spacings) or to a certain structural arrangement. An example is the stacking of layers during growth, so that different sequences of, for example, fcc and hcp stacking can be clearly differentiated.

Complex crystallographic structures are unavoidable when there is competition between pseudomorphic growth and a film's tendency to assume its native structure. Then, substantial distortions can develop within the film as, for example, in the case of iron on reconstructed iridium, where the substrate's structure influences the film growth in an unusual way. The power of LEED is demonstrated for these scenarios also.

Magnetic films deposited on single-crystal substrates have been the subject of numerous studies since their bi-dimensional or nanostructured character provides important extensions of the family of bulk magnetic solids. For these objects, the knowledge of their size, shape and structure at the atomic scale are mandatory for an understanding of their unusual magnetic properties. After a short introduction to the strengths of extended x-ray absorption fine structure (EXAFS) for local structure determination, we will concentrate on three well defined cases. Thin Ni layers deposited on Cu(001) present a perpendicular magnetic anisotropy in a very wide thickness range, and thin Fe layers on MgO(001) cut into stripes by the 'atomic saw' (AS) method have a magnetization easy axis oriented perpendicular to the stripes. In both cases, a precise structural characterization obtained by EXAFS and simple magneto-elastic models allow description of their magnetic behaviour. The growth of a cobalt film on a vicinal Cu(11n) surface induces a uniaxial magnetic anisotropy. Although the steps are clearly at the origin of this behaviour, the exact mechanism (lower coordination at step edges, anisotropic strain) remains obscure. Detailed studies of the morphology by STM and of the structure by EXAFS allow the conclusion that the step-induced magnetic anisotropy does not originate from an in-plane anisotropic structural relaxation but is more likely related to the stepped film morphology.

Scanning tunnelling spectroscopy (STS) of thin Fe films on W(110) shows that the electronic structures of magnetic domains and domain walls are different. This experimental result is explained on the basis of first-principles calculations. A detailed analysis reveals that the spin–orbit induced mixing between minority d_xy+xz and minority d_z² spin states depends on the magnetization direction and changes the local density of states in the vacuum detectable by STS. The effect scales in second or fourth order with the magnetization angle relative to the easy axis. Our finding implies that nanometre-scale magnetic structure information can be obtained even by using non-magnetic probe tips. Magnetization dependent measurements show that the canting of adjacent spins has no major influence on the electronic structure of the sample.

By use of photoemission spectroscopy with high resolution in both energy and momentum, one is able to observe a considerable splitting of the Shockley-type surface state on the (111)-face of gold. This splitting exists over the complete two-dimensional Fermi surface of this surface state but disappears at normal emission. A comparison with fully relativistic density functional calculations proves that the splitting is caused by spin–orbit interaction, which—in combination with the lack of inversion symmetry at the surface—resolves the spin degeneracy of the surface state band. This paper gives a summary of recent experimental and theoretical investigations on the split Shockley state on Au(111). In addition, it will be demonstrated that the splitting is influenced by adsorbates, and it will be discussed why it has not been observed experimentally on other noble metals.

The interaction of photons with spins is mediated through spin-orbit coupling. Therefore, magneto-optics of magnetic materials depends crucially on the spin-orbit interactions in either the initial and/or the final state of the optical transition considered and this makes magneto-optics a very powerful tool for the study of spin-orbit coupling in 3d, 4f and 5f materials. This paper will review some of the most striking results obtained over the last 30 years (For the data up to 1990 see: Schoenes J 1991 Materials Science and Technology Vol 3A Electronic and Magnetic Properties of Metals and Ceramics eds. R W Cahn, P Haasen and E J Kramer (Weinheim: VCH)p147).

The largest ever reported Faraday rotations exceed 2×10 ⁶ deg cm^-1. They have been observed in thin films of the europium monochalcogenides EuS, EuSe and EuTe, which have a 4f ⁷ ( ⁸S_7/2) groundstate, i.e. no spin-orbit splitting in the initial state. The occurrence of the large magneto-optical signal is therefore clear evidence for a large spin-orbit interaction in the final state of the responsible transition. A line shape analysis shows that this final state consists of a coupling of the excited 5d electron with the remaining 4f ⁶ state. Nd₃S₄ is an example of a system with spin-orbit coupling in the initial as well as in the final state. The diamagnetic line shape helps to identify the coupling in the final state and the spin-orbit parameter can be determined.

While the intrinsic value of the Kerr rotation of CeSb is still debated, agreement exists that in any case CeSb has the largest Kerr rotation. Of particular interest is the transition from a paramagnetic line shape to a diamagnetic line shape, when the magnetization is reduced from saturation. LaSe presents another archetypical example. Here the initial state is a non-magnetic 5d¹ state and the magneto-optical experiment allows one to study empty 4f states.

In all these lanthanide compounds Russel-Saunders coupling could be applied to the initial and final states. This is no longer the case for the final state in actinide compounds. Magnetic circular dichroism and Faraday rotation measurements on UO₂ films evidence that the final state for the excitation of an f electron out of the 5f ² initial state is near to a j-j coupled 5f ¹-6d¹ state. A different approach has to be made to interpret magneto-optical data for uranium monochalcogenides. In the case of US, for example, 5f delocalization and f-d hybridizations are too large to apply atomic models and sophisticated band structure calculations are needed to account for the spectra.

Besides interband transitions, magneto-optics has also proven to be powerful for studying free carriers. Two different types of frequency dependencies have been predicted by theory. Examples of these two kinds will be presented in the form of data for TmSe and U₃P₄. The special case of CrPt₃ will be discussed in a separate paper.

Spin–orbit coupling (SOC) plays a major role in solid-state magnetism. Its effects range from completely defining the splitting of f states in rare earths to the correct prediction of the ground-state orientation of the magnetic moment in transition metals. The breaking of symmetries that it produces is the main tool for addressing magneto-optical effects. In this paper we theoretically study the possibility of all-optical spin switching. By simulating the effect of an ultrashort laser pulse on an electronic system, we show the crucial role of SOC in the spin dynamics occurring in this regime. The role of other parameters is also analysed. Finally, we show the possibility of obtaining an all-optical spin switching in the subpicosecond time regime.

In this paper we discuss various contributions to the transient magneto-optical (MO) response of ferromagnetic metals after laser-heating by a femtosecond pump pulse. It is demonstrated that a simple relation between MO response and the fast magnetization dynamics does not hold. A more general concept is introduced, on the basis of which artifacts are experimentally identified. Special focus is on problems associated with conservation of angular momentum in the ultrafast regime. A model describing dichroic state filling effects, fully including the coherent interactions between pump and probe pulses, is worked out. One of the potential methods to achieve access to the genuine magnetization dynamics, based on a thermal difference scheme, is analysed in detail. Finally, it is discussed how the different dynamical processes triggered by the laser pulse appear in the transient MO signal.

Klaus Baberschke 2002

The 281 Wilhelm und Else Heraeus Seminar on the occasion of the 65th birthday of Klaus Baberschke

In December 2001 Klaus Baberschke celebrated his 65th birthday. In connection with this occasion a four day scientific workshop was held at Wandlitz near Berlin as well as a colloquium at the Freie Universität Berlin. Klaus Baberschke has worked and taught successfully at the FUB for more than 30 years and still continues to do so. 24 invited and 35 contributed papers were presented on subjects related to the role of spin--orbit coupling in condensed matter physics, a topic to which Klaus Baberschke has made many outstanding contributions, starting from g-factors, to crystal field effects and to recent applications of XMCD sum rules. The papers in this special volume reflect some of the many interests of Klaus and also his attitude towards research: high quality and clarity of presentation. They give an outstanding summary of the current state of the art in theory and experiment in this special field of magnetic properties in low dimensional systems. The truly international workshop participation also testifies to the strong scientific links Klaus has enjoyed across the world, an important ingredient for high quality research in the fast evolving field of low dimensional magnetism. The organizers have benefited from many years of successful collaboration and illuminating discussions with Klaus. The work will go on.

Thank you, Klaus.

Many magnetic phenomena in nanoscale structures are strongly influenced by the interaction of orbital magnetic moments and spin moments. For example, anisotropic magnetoresistance, magnetostriction, the magnitude and anisotropy of magnetic moments, and the overall magnetic anisotropy energy of ferromagnets are intrinsically related to this interaction. In today's low dimensional magnetic structures, spin-orbit influenced phenomena become even more important, since the well known quenching of the orbital magnetic momentum encountered in bulk systems is lifted. Furthermore, magnetic systems may be prepared with highly strained crystallographic structures in phases that do not occur naturally in the bulk. In order to characterize the real space and magnetic structure of such novel low dimensional magnets, x-ray, neutron and electron based techniques are commonly used. Very often discrepancies are found in the results not only with theory but also between various experimental probes, highlighting the difficulty of the problem and the need for a comprehensive dialogue between experimentalists and theoreticians.

The purpose of this 281 Wilhelm und Else Heraeus Seminar was the gathering of leading experts from different fields of solid state physics to discuss the different approaches of calculating and measuring spin-orbit influenced phenomena in magnetic low dimensional systems and nanostructures. 70 participants from Germany, France, Italy, Japan, Poland, Romania, Sweden, Ukraine and the USA attended the four day workshop and heard 24 invited lectures and saw 35 poster presentations. The results of the lectures and lively discussions which ended often late at night are summarized in the contributions to this special volume. We feel that the authors have accomplished the difficult task of combining a tutorial type of introduction to their specific research area with an excellent presentation of their latest results. The interested reader will find contributions discussing the results obtained by a rich variety of experimental techniques such as photoemission, magnetic and natural x-ray dichroism, paramagnetic and ferromagnetic resonance, neutron scattering and reflectometry, magneto-optics and femtosecond spectroscopy. The numerous experimental results covering aspects of the g-tensor analysis in magnetic monolayers, exchange-coupling, spin-orbit splitting in electronic band structure and excited crystal field states were complemented by state of the art theoretical presentations of, for example, the magnetic susceptibility and orbital and spin magnetic moments obtained from ab-initio calculations.

We would like to thank the authors for their excellent contributions and the chairpersons and participants for making the 281 Wilhelm und Else Heraeus Seminar such a high-level event in terms of scientific output and lively interactions.

The programme of the seminar is available below as a pdf file.