to the main page
	webmaster mail

Scientific background

These novel developments have primarily involved semiconductors, but more recently have been applied to metals: their growth by Molecular Beam Epitaxy (or other similar techniques) allows a control of the thickness at the atomic layer scale. Metals are important mainly for their possible magnetic properties and applications as storage media and sensors.

This field is currently very active, not only for the possible applications of magnetic nanostructures. In fact, metallic films are potentially important toward revealing fundamental aspects of low-dimensional systems.

In addition, the growth process itself is extremely interesting because the study of the dynamics of a growing surface is a classic topic of nonequilibrium statistical mechanics: it implies kinetic roughening and growth instabilities.

The growth process is clearly relevant for determining the magnetic properties of the nanostructures: generally, sharp interfaces and no interdiffusion are required, but in some cases the growth can be governed in order to produce self-organized structures, for example step decoration on vicinal surfaces or magnetic dots.

Nowadays experimenters are studying smaller and smaller structures to reach the nanoscale, while the atomic simulations are possible for larger and larger systems. For the artificial structures we are considering, experiences and numerical modelings are coming to a meeting point. The tools used to study the elementary processes that drive the growth and the surface morphology are molecular dynamics, Monte Carlo with atomic displacements and energy minimum path calculations. The Kinetic Monte Carlo method with particles diffusing on a crystal surface, as for instance in the solid-on-solid models, is used to determined island shapes and sizes depending on the particles types and their deposition conditions. The chemical profile, which is also a very important factor for the magnetization at the surface, can also be studied. One of the actual goals of the simulation is to integrate the most pertinent information calculated from the small length and time scales simulations into these Kinetic Monte Carlo.

Simulation of magnetic structures are also performed i) at the atomic level or ii) at a much larger scale. In the first approach, magnetic configurations are found by tight-binding or ab-initio calculations for different chemical profiles at the surface like, for instance, Fe on Cu, Ni on Cu, or Cr on Fe. The second approach does not consider atoms, but the magnetization of small cells of the material. Thus an Hamiltonian for Ising spins or vector spins is introduced and Monte Carlo simulations are performed to investigate the complex magnetic domain structures.

Motivation and objectives

The same situation is true for the physicists doing numerical simulations. It is now important to have a view of the present bridges between growth studies and magnetic ones, and to investigate what are the theoretical difficulties which have to be overcome.

We think that improving the contact between the two communities is essential for both fundamental and technological developments. It is the main aim of this Workshop to make an attempt in this direction.

Furthermore, we would like to improve the contact between experiments and simulations, as we think that both could find fruitful new directions by listening to their respective needs.

Indeed, the computational approach, which is well designed to study nanostructures, can give information that is not accessible through experiments but is essential to understand the link between growth and magnetism and to predict materials properties.

Discussion

1 ) Do magnetic properties influence the growth process?

This is a very fundamental question and a very delicate one. From the theoretical point of view, the influence of magnetism on growth has not been worked out on a microscopic level. Intuitively, the energies involved in magnetism are much smaller than the ones relevant in crystal growth, so that one does not expect that magnetism alters essentially the growth process. On the other hand, one really would like to see more clearly on the role of magnetostriction in modifying the growth process.

2 ) What is the role of morphology (steps, defects and roughness) in determining the magnetic response?

Morphology has often been taken into account in an `effective' way, through the introduction of new terms in the Hamiltonian describing the magnetic system. For example, steps on a vicinal surface introduce additional anisotropy terms; roughness of the interfaces in a superlattice may give rise to the biquadratic exchange between magnetic films, or to anisotropies favoring a canting of the magnetization. Are there cases where the effect of this kind of `disorder' is more complex?

3 ) How does magnetism develop in the initial stage of growth?

The growth in the submonolayer regime has some similarities with a percolation process, but there is a major difference: adatoms diffuse at the surface. Furthermore, long-range forces (mainly dipolar interaction) may strongly modify the magnetic properties, and induce a long-range order even before the percolating cluster appears.

4 ) Can we manipulate magnetic properties by ad hoc atomic scale manipulation?

This is the central question which will involve researchers for a long time to go. Both is available: accurate inspection of magnetic properties and atomic scale manipulation. Integrating them into a process would give rise to immense new possibilities, from both fundamental and technological point of view. One dreams too often of being able to switch magnetization by controlled adding or subtracting of --say-- single atoms!

Paolo POLITI
  Danilo PESCIA
Frédéric LANÇON