1. INTRODUCTION

The design and fabrication of magnetic nanostructures and the study of their specific properties are playing an increasingly important role directly related with the large number of potential applications in various fields such as information storage and magnetoelectronic devices. ^{1 5} From the fundamental point of view, the magnetic behaviour of individual, isolated nanoscale systems was mainly theoretically approached while experimental investigations concerned essentially nanoparticles assemblies because of the sensitivity limitation of the magnetic characterization techniques. On the theoretical side ^[1], ab-initio calculations within the now standard Local Spin Density (LSD) approximation to the Density Functional Theory (DFT) have proven to be a valuable tool to study the structure and electronic properties of systems up to some few hundreds of atoms. In particular, the spin order of nanosize systems (atomic monolayers, small clusters, etc) is now within reach of such calculations. In the last few years, the appearance of non-colinear spin functionnals6 combined with attempts to provide a framework for molecular dynamics on the spin degrees of freedom7 ^,8 allowed to extend the field of application of such approaches. It remains nevertheless that the energy range at which e.g. spin-orbit effects occur (i.e. a few meV) is still a challenge for ab-initio calculations aiming at calculating for example the barrier (the anisotropy) for spin flip in nanosize systems. On the experimental side, the recent...