$\textit{Ab initio}$ derivation of effective Hamiltonian for La$_2$CuO$_4$/La$_{1.55}$Sr$_{0.45}$CuO$_4$ heterostructure

Abstract

We formulate a method of deriving effective low-energy Hamiltonian for nonperiodic systems such as interfaces for strongly correlated electron systems by extending conventional multi-scale $\textit{ab initio}$ scheme for correlated electrons (MACE). We apply the formalism to copper-oxide high $T_c$ superconductors in an example of the interface between overdoped La$_{2-x}$Sr$_x$CuO$_4$ and Mott insulating La$_2$CuO$_4$ recently realized experimentally. We show that the parameters of the $E_g$ Hamiltonian derived for the La$_2$CuO$_4$ /La1.55Sr0.45CuO$_4$ superlattice differ considerably from those for the bulk La$_2$CuO$_4$, particularly significant in the partially-screened Coulomb parameters and the level offset between the $d_x^2-y^2$ and $d_z^2$ orbitals, $\Delta E$. In addition, we investigate the effect of the lattice relaxation on the $E_g$ Hamiltonian by carefully comparing the parameters derived before and after the structure optimization. We find that the CuO$_6$ octahedra distort after the relaxation as a consequence of the Madelung potential difference between the insulator and metal sides, by which the layer dependence of the hopping and Coulomb parameters becomes more gradual than the unrelaxed case. Furthermore, the structure relaxation dramatically changes the $\Delta E$ value and the occupation number at the interface. This study not only evidences the importance of the ionic relaxation around interfaces but also provides a set of layer-dependent parameters of the $E_g$ Hamiltonian, which is expected to provide further insight into the interfacial superconductivity when solved with low-energy solvers.

Terumasa Tadano

Researcher of Materials Science

My research interests include development of computational methods and softwares for predicting thermal properties of solids, and application of machine-learning methods to material science study