International audienceThis review summarizes recent first-principles investigations of the electronic structure and magnetism of dilute magnetic semiconductors (DMSs), which are interesting for applications in spintronics. Details of the electronic structure of transition-metal-doped III-V and II-VI semiconductors are described, especially how the electronic structure couples to the magnetic properties of an impurity. In addition, the underlying mechanism of the ferromagnetism in DMSs is investigated from the electronic structure point of view in order to establish a unified picture that explains the chemical trend of the magnetism in DMSs. Recent efforts to fabricate high-TC DMSs require accurate materials design and reliable TC predictions for the DMSs. In this connection, a hybrid method (ab initio calculations of effective exchange interactions coupled to Monte Carlo simulations for the thermal properties) is discussed as a practical method for calculating the Curie temperature of DMSs. The calculated ordering temperatures for various DMS systems are discussed, and the usefulness of the method is demonstrated. Moreover, in order to include all the complexity in the fabrication process of DMSs into advanced materials design, spinodal decomposition in DMSs is simulated and we try to assess the effect of inhomogeneity in them. Finally, recent works on first-principles theory of transport properties of DMSs are reviewed. The discussion is mainly based on electronic structure theory within the local-density approximation to density-functional theory

We calculate the persistent current of 1D rings of spinless fermions with shortrange interactions on a lattice with up to 20 sites, and in the presence of disorder, for various band fillings. We find that both disorder and interactions always decrease the persistent current by localizing the electrons. Away from half-filling, the interaction has a much stronger influence in the presence of disorder than in the pure case.

We present a semi-analytic theory for the Curie temperature in diluted magnetic semi-conductors that treats disorder effects exactly in the effective Heisenberg Hamiltonian, and spin fluctuations within a local RPA. The exchange couplings are taken from concentration dependentab initio estimates. The theory gives very good agreement with published data for wellannealed samples of Mn x Ga 1−x As. We predict the critical temperatures for Mn x Ga 1−x N lower than in doped GaAs, despite the stronger nearestneighbour ferromagnetic coupling. We also predict the dependence on the hole concentration.Search for diluted magnetic semiconductors with ferromagnetism stable to high temperatures has been hampered by the many physical parameters that may determine magnetic properties. These include the choice of host semiconductor, that of the doping magnetic impurity, the degree of compensation, and methods of preparation and treatment of the sample [1]. The underlying mechanism of interaction between dopant spins has without doubt been correctly identified: Ruderman-Kittel-Kasuya-Yosida(RKKY)-like effective interactions mediated by both the host [2,3] and the doping band [4-6]. Despite this, the theory has not lead to reliable quantitative predictions. Comparison of calculations has been complicated by difficulties of full characterizing the samples. In samples of GaAs doped with Mn, extensive experimental studies [1] have now allowed for greater control over sample parameters and there is now apparently convergence to reliable experimental values of the critical temperature of different groups [7-10]. An important factor was that the carrier densities were measured simultaneously by magneto-transport. There is now the possibility of testing calculations against experiments and determining the origin of past discrepancies.Ab initio calculations using the Local Density Approximation and the magnetic force theorem can be used [5] to derive realistic values of magnetic exchange interactions between classical impurity spins. These calculations also take into account the effect on the effective exchange of disorder of the carriers, within a Coherent Potential Approximation (CPA). Similar calculations based on supercells [12,13] lead to comparable results at low concentration. It has become apparent that the difficulty is not in deriving the effective magnetic Hamiltonian correctly but in treating its thermodynamics accurately. As we will demonstrate explicitly here, treating the magnetic correlations by over-simplified mean field theories [14,15] has lead to overstatement , by a wide margin, of the critical temperature T c . The disorder in the effective magnetic model also plays an important role that cannot be simply treated by an effective medium theory of the style of the Virtual Crystal Approximation (VCA) [16]. This suggests that the discrepancy with experiment is largely due to approximations made to the effective Hamiltonian, not the values of the couplings themselves. Thus by improving the treatment of the effective Hamilt...