The development of spintronics has stimulated the study of ferromagnetic thin films, used for magnetic data recording and reading. Although the GMR effect has been demonstrated in Fe/Cr superlattices, antiferromagnetic thin films only play an auxiliary role for these devices. Recent theoretic studies however proposed they could play a more active role, replacing ferromagnetic thin films, with competitive and distinct properties. Contrary to ferromagnetic layers, antiferromagnetic thin films have been much less experimentally examined because specific techniques (like neutron diffraction) are needed to that purpose, beyond usual characterization means. During this thesis work, we examined a model system for spintronics consisting in a thin film of chromium (001), an archetypal itinerant antiferromagnet, and a (001) magnesium oxide layer, widely used as a crystalline tunnel barrier in spintronic devices. We have demonstrated the possibility to control the magnetic anisotropy and the magnetic order of the antiferromagnet by mastering the growth conditions, and by doping Cr with Fe. Thanks to a detailed photoemission study, we have extended proofs of the interplay between magnetism and electronic structure in Cr, and in particular, we have evidenced surface and interface polarized states. By completing these results with neutron reflectivity experiments, we determined the magnitude of the ferromagnetic moments at the Cr/MgO(001) interface. Finally, by exploiting this gathered amount of knowledge, we succeeded in understanding and interpreting the characteristic magnetism observed in Cr/MgO heterostructures with thin tunnel barriers as the sign of a tunnel interlayer coupling.
