Spin-polarized electron transport across metal-organic molecules: a density functional theory approach

  • Author:

    A. Bagrets

  • Source:

    J. Chem. Theory Comput., 2013, vol.9 (6), pp.2801-2815

  • Date: 2013
  • In the field of molecular spintronics, experimental techniques have achieved a stage where it is feasible to explore the interplay between quantum electron transport and magnetism at the single molecule level. An example is a spin-polarized STM, which can probe local electrical currents through organic molecules deposited on magnetic surfaces. The atomistic complexity of nanoscale systems calls for a first-principles description of spin-dependent transport phenomena, e.g., based on the non-equilibrium Green's function (NEGF) formalism merged with density functional theory (DFT). However, for the case of molecular junctions with transition metal electrodes, a computation of the underlying Kohn-Sham Hamiltonian can be a challenging problem: a simultaneous and accurate description of spin ordered magnetic surfaces together with the electronic structure of a molecule is required. In the present work, we provide a solution for this problem. We present an implementation, within a standard quantum chemistry package, of the NEGF formalism with an efficient approximation for the self-energy, which accounts both for absorbing boundary conditions and for exchange splitting of the energy bands in ferromagnetic electrodes. We demonstrate an ability to simulate a variety of magnetic configurations including nanoscale domain walls, which are realized when a molecule with few spin centers is brought in contact with differently magnetized reservoirs. Magnetoresistance effect arising at molecular scale is discussed based on examples including Ni atomic-sized contact, electron transport across a prototypical molecular magnet (vanadium-benzene multidecker cluster) and tunneling through Co-phthalocyanine. Furthermore, we verify stability of magnetically nontrivial solutions against electron correlations within the DFT+U approach.