Friday, December 14, 2012

1212.2976 (Shiu Liu et al.)

Impurity-induced frustration: low-energy model of diluted oxides    [PDF]

Shiu Liu, A. L. Chernyshev
We provide a detailed derivation of the low-energy model for Zn-diluted La2CuO4 in the limit of low doping together with a study of the ground-state properties of that model. We consider Zn-doped La2CuO4 within a framework of the three-band Hubbard model, which closely describes high-Tc cuprates on the energy scale of the most relevant atomic orbitals. Qualitatively, we find that the hybridization of zinc and oxygen orbitals can result in an impurity state with the energy \varepsilon, which is lower than the effective Hubbard gap U. The low-energy, spin-only Hamiltonian includes terms of the order t^2/U and t^4/\varepsilon^3. That is, besides the usual nearest-neighbor superexchange J-terms of order t^2/U, the low-energy model contains impurity-mediated, further-neighbor frustrating interactions among the Cu spins surrounding Zn-sites in an otherwise unfrustrated antiferromagnetic background. These terms can be substantial when \varepsilon ~ U/2, the latter value corresponding to the realistic CuO2 parameters. In order to verify this spin-only model, we subsequently apply the T-matrix approach to study the effect of impurities on the antiferromagnetic order parameter. Previous theoretical studies, which include only the dilution effect of impurities, show a large discrepancy with experimental data in the doping dependence of the staggered magnetization at low doping. We demonstrate that this discrepancy is eliminated by including impurity-induced frustrations into the effective spin model with realistic CuO2 parameters. Recent experimental study shows a significantly stronger suppression of spin stiffness in the case of Zn-doped La2CuO4 compared to the Mg-doped case and thus gives a strong support to our theory. We argue that the proposed impurity-induced frustrations should be important in other strongly correlated oxides and charge-transfer insulators.
View original: http://arxiv.org/abs/1212.2976

No comments:

Post a Comment