Computational Condensed Matter Physics Laboratory

Outline

Our main aim is to theoretically understand various novel quantum phases and phenomena in a wide range of materials by microscopically studying electronic structures. Our main interests include strongly interacting electronic systems such as different kinds of transition metal oxides (cuprates, manganites, etc.) and low-dimensional organic compounds, for which novel quantum states, low-lying collective excitations and quantum transport properties are studied using various state-of-the-art numerical methods. Oxide heterostructures, mainly based on transition metal oxides, is one of our recent focused projects to theoretically propose a new functionality for strongly correlated electronic devices. We also focus on searching for novel quantum states of systems with a large spin-orbit interaction such as 5d transition metal oxides. We are also devoted to developing new numerical methods for quantum many-body systems in general.

Recent Research Topic

Quantum simulations for a spin-orbit-induced Mott insulator

Fig. 1 Ground state phase diagram and one-particle excitation spectra

(a) The ground state phase diagram as functions of the on-site Coulomb interaction U/t and the spin-orbit coupling λ/t, and (b) the single-particle excitation spectrum (left) and the projected single-particle excitation spectrum onto Jeff=1/2 (right) for the insulating phase.

Recent resonant X-ray scattering experiments on Sr2IrO4 have revealed that the low-lying single particle excitations in the upper Hubbard band consist of electronic states with dxy:dyz:dzx=1:1:1, which indicates that the electron-unoccupied states of the lowest single-particle excited states are described by an effective angular momentum Jeff=S-L=1/2, and thus this insulating state is called a spin-orbit-induced Mott insulator. To understand the ground state properties as well as the low-lying excitations of Sr2IrO4, we have constructed an effective three-band Hubbard model on the square lattice. Using the variational Monte Carlo study, we first established the ground state phase diagram as functions of the spin-orbit interaction l/t and the local Coulomb interaction U/t (Fig. 1 (a)). We found that with an appreciable value of U/t, the spin-orbit interaction can induce a new type of Mott insulator (spin-orbit-induced Mott insulator), where the Mott insulator-metal transition occurs by the smaller critical value of U/t when the spin-orbit interaction is strong. To understand the nature of this insulating state, we have employed a variational cluster approximation and calculated the single-particle excitation spectra. The obtained results are shown in Fig. 1 (b) for the Mott insulating phase. We can clearly see that the lowest-lying excited states close to the Fermi level have the characteristic of Jeff=1/2, although there seems to exist mixed states of Jeff=1/2 and Jeff=3/2 for the electron removal spectra in a small region of momenta. Our result offers direct numerical evidence that the spin-orbit-induced Mott insulator appears in this system.

Anomalous enhancement of the spin Hall conductivity in a superconductor/normal metal junction

We have proposed a spin Hall device to induce the large spin Hall effect in a superconductor/normal metal (SN) junction (Fig. 2). The side jump and the skew scattering contributions have both been taken into account to calculate the extrinsic spin Hall conductivity in the normal metal. We have found that both contributions are anomalously enhanced when the voltage between the superconductor and the normal metal approaches the superconducting gap. This enhancement is attributed to the resonant increase of the density of states in the normal metal at the Fermi level. Our results have clearly demonstrated a novel way to control and amplify the spin Hall conductivity by applying an external dc electric field, suggesting that SN junctions have a potential application for a spintronic device with a large spin Hall effect.

A Haldane-Anderson impurity model study for the spin- and charge-states of iron in heme proteins

To understand the spin and charge properties of heme proteins such as myoglobin, we have formulated a Haldane-Anderson impurity model using first-principles quantum chemical calculations based on the density functional theory with the closed-shell condition and with a hybrid-GGA approximation. First, we have employed a mean-field approximation to solve a simple model of myoglobin active site, the iron porphyin-imidazol complex FeP(Im), and found that the results reproduce some of the known spin-charge states of myoglobin. We have computed the spin-charge phase diagram of iron in FeP(Im) with and without O2 attached to the central iron (Fig. 3), and found that the spin-charge states for FeP(Im) are much more sensitive to the local correlations than that for FeP(Im)(O2). We are planning to use the continuous-time quantum Monte Carlo method and the density matrix renormalized group method to understand the correlation effects correctly on the spin-charge state of iron in heme proteins.

Fig. 2 A schematic configuration of a superconductor/normal metal junction proposed to induce the large spin Hall effect

V is the voltage applied between superconductor (S) and normal metal (N). Vbias is an applied dc voltage at opposite edges of the N. An insulating barrier (I) is introduced between S and N.

Fig. 3 Spin-charge states of iron in myoglobin

Spin-charge states of iron in FeP(Im) (in black) and FeP(Im)(O2) (in grey). The lines indicate the boundaries of each phase, whose nature is indicated in the delimited area. U (direct) and J (exchange) are local Coulomb interactions.