Research

Overall research area of our group is computational materials science, and interdisciplinary between solid state physics and physical chemistry. Our main research topics are described in the following.

1. Two-dimensional magnets

Background Artificial intelligence or data-based computing is growing fast these days. Conventional (the von Neumann) architecture suffers from data transfer bottleneck between processor and memory, and a new architecture enabling the data-intensive computing is needed such as process-in-memory (PIM). Among various PIM candidates, spin-based system is promising because of low-energy operation, non-volatile storage and compatibility with established magnetic memory technology. Developing novel spintronic materials towards PIM requires a lot of experimental efforts and costs. On the other hand, theoretical studies can help not only to screen choices of magnetic materials, but also to find novel quantum materials and fundamental physics, which will be useful for next-generation information technology. Our group studies magnetic properties of emerging two-dimensional magnets using first-principles calculations towards novel vdW spintronic devices.

(i) van der Waals (vdW) multiferroics

The experimental discoveries of 2D magnetism provide many opportunities to explore various types of heterolayers towards vdW spintronics as there is no restriction of lattice matching. For an electrical control of the 2D magnetism, we theoretically studied a hetero-bilayer of 2D ferroelectric (FE) In₂Se₃ and 2D ferromagnetic (FM) Cr₂Ge₂Te₆ [1]. According to the Mermin-Wagner theorem, a long-range magnetic ordering in 2D system is allowed only for the Ising-type spins with the magnetic anisotropy along the out-of-plane direction. This fundamental property is exploited to switch the favored magnetic direction of Cr₂Ge₂Te₆ between in-plane or out-of-plane by the upward or downward electric polarization, respectively, of the In₂Se­₃. From the detailed electronic structure analysis, it has been found that the interlayer orbital hybridization is either significant or negligible depending on the electric polarization direction of In₂Se₃, and the hybridization strength varies large enough to switch the magnetocrystalline anisotropy (MCA), thus electrically switchable net magnetic moment of Cr₂Ge₂Te₆ has been proposed. In addition, the induced intralayer multiferroicity in In₂Se₃ could open a new concept of magnetoelectric devices, using a single knob (direction of electric polarization in In₂Se₃) to control the magnetic order in both In₂Se₃ and Cr₂Ge₂Te₆.

(ii) interlayer magnetic coupling

Controlling the interlayer magnetic coupling in the 2D layered magnets holds great potentials in both manipulating the ground-state magnetic states and engineering the layered structures for the cross-layer electron transport or tunneling phenomena. For example, the CrI₃ has shown the stacking-type dependent ferromagnetic (FM) or antiferromagnetic (AFM) coupling between adjacent layers. From many theoretical studies, the carrier doping and pressure (or interlayer distance) are likely to influence significantly the interlayer exchange coupling (IEC). Also the interplay between the spin-orbit coupling and the interfacial symmetry deserves fundamental studies towards intriguing magnetic states.

Recently we have studied the IEC of metallic Fe₃GeTe₂ (FGT) which is famous for the room-temperature 2D magnets together with Fe₄GeTe₂ and Fe₅GeTe₂. Reported experiments suggest coexistence of FM and AFM IEC couplings for bulk FGT and the hard magnetic behavior for a thin film. To understand this behavior, we studied the stacking type and thickness effects on the IEC strength. Interestingly, we found the enhanced FM interlayer coupling with lowering the stacking symmetry as well as introducing the surface boundary [2]. From the small amount of stacking-fault energy that we found as well, our study suggests that the coexisting AFM IEC coupling or the short FM correlation length would originate from the electronic structure itself rather than the stacking fault. Also a knob of the non-local screening is suggested to control the IEC strength.

2. Heterojunction electronic structure

Background Overall performance in various types of optoelectronic and transistor devices is not only given by the intrinsic properties, but also significantly influenced by the electronic coupling at the heterojunction. For example, depending on the band alignment and/or electron redistribution, the electron transport behavior across the interface can be characterized by Ohmic or Schottky contact. For the heterojunction electronic structure, a full ab initio description is still computationally unfeasible due to the extended system size, and an efficient calculation based on the density function theory (DFT) is prevailing. By using the DFT method and sophisticated treatments, our group studied electronic structure of hybrid perovskite and its heterojunction with carrier extraction materials. In combination with the non-equilibrium Green’s function method, we designed an efficient electron or spin transport channel material based on the graphene nanoribbons. Our studies provide microscopic details of heterojucntion with helpful guidance towards high-performance devices.

Hybrid perovskite

(i) polaron transport in the hybrid perovskite

Despites of substantial amount of defects to be present for the solution processing, the hybrid perovskite solar cell (PSC) exhibits the power conversion efficiency (~25%) comparable to that of the Si-based solar cell. Besides a proper band gap size and a high absorption coefficient, the main reasons are attributed to the defect-tolerant electronic structure and high carrier mobility (10¹~10³ cm²/Vs). Regarding the high carrier mobility, the main origin has been ascribed to an intriguing coupling behavior between a charge carrier and the lattice vibration, or polaron quasiparticles. Our group studied the role of organic cations on the transport behavior of Frohlich (large) polarons [3]. For various types of organic cations, the carrier scattering rates have been investigated, and a new perspective of controlling the coupling strength between organic cations and halide ions has been suggested to enhance the carrier mobility.

(ii) heterojunction electronic structure in the hybrid perovskite solar cell

Within a solar cell device, the electron transport material (ETM) and hole transport material (HTM) are incorporated alongside the halide perovskite to facilitate the charge (electron and hole) separation process at the respective interfaces. Therefore, a comprehensive understanding of the E(H)TM/perovskite interface is crucial for elucidating the electron (hole) transfer process across the interface and its influence on the photovoltaic properties. It was known that for bulk hybrid perovskite the Rashba-Dresselhaus (RD) effect is one of reasons responsible for suppressed recombination rate of carriers. Nevertheless, the heterojunction is likely to have the enhanced RD effect due to the asymmetric structure. Our group reported a sizable or robust Rashba-Dresselhaus effect at the interface between perovskite and ETM (TiO₂) [ACS Energy Lett. 3, 1294 (2018)], by this the suppressed electron-hole recombination could facilitate electron transfer to ETM.

The band alignment at the heterojunction is the most elementary and important property to judge the carrier extraction characteristics. From our study on the interface electronic structure between TiO₂ and MAI-terminated MAPbI₃, an interesting role of proton transfer or a short hydrogen bonding on the band alignment has been found [J. Mater. Chem. A 6, 4305 (2018)]. A proton from the ammonium can reside very close to an O atom of TiO₂, thus the electron transfer becomes favorable. Our work suggests a beneficial role of the hydrogen-bond reaction at the perovskite-oxide interface on the electron transfer.

For practical commercialization of PSCs, the long-term stability and large area growth should be resolved. The presence of organic cations causes the perovskite unstable against humidity. Among various attempts made to protect the PSCs from water vapors, one approach has been to introduce an interlayer with hydrophobic property, for example graphene-based materials. Our group studied the heterojunction electronic structure between MAPbI₃ and a functionalized graphene or graphene nanoribbon to secure a selective extraction or blockage of electron or hole carriers [J. Mater. Chem. A 6, 18635 (2018)]. While the interface band alignment within DFT is commonly predicted by comparing the bulk electronic structures of individual components with the electrostatic reference potential, a significant effect of the interface coupling or surface boundary is considered for the band alignment [Phys. Chem. Chem. Phys. 22, 2955 (2020) ]. It motivates further theoretical method development for proper description of the non-local screening and quasiparticle effects.

Efficient electron/spin channel material

(i) transport gaps in zigzag-edge graphene nanoribbons (zGNRs) with chemical edge disorder

Graphene is very attractive for nanoscale electronics owing to high carrier mobility as well as high mechanical strength and thermal conductivity. It can be used as an atomically thin channel material for low-power field-effect transistors with a proper procedure for band gap opening like the width confinement as in GNRs. But fabricated devices could not reach a sufficient current on/off ratio, greater than ~10⁴, due to substantial gap states caused by the edge roughness. The zGNRs exhibits peculiar edge-localized states near the Fermi level, thus heterogeneous chemical modification along the edges can be used to open the transport gap. Our group studied the emergence of transport gap in zGNRs with chemical edge disorder [Applied Surface Science 512, 144714 (2020) ]. It has been found that the narrow width is helpful to enhance the on/off ratio as the transport gap starts to appear more abruptly with the disorder strength due to the orthogonal property between two bulk transport modes.

(ii) spin filtering in defective zGNRs

The zGNRs attract a revived attention owing to atomically straight edge shape realized experimentally. The half-metallicity under a transverse electric field has been predicted theoretically, but a new facile approach without external field is interesting for 1D spin channel with a superior spin relaxation time from the d⁰ magnetism. It is known that pristine zGNRs have a short spin correlation length and suffer from a significant spin fluctuation due to quasi-degeneracy of different magnetic states. Our group designed a zGNR with the nitrogenated divacancy framework and showed that various kinds of an adatom at the framework causes the half-metallic ferromagnetic state to be most stable [Nanoscale Advances 2, 5905 (2020)]. The key mechanism for the stable ferromagnetic ground state is the sufficient electron doping by the adatom with preserving the inherent magnetic property of the zigzag-edge defect. In addition, the most stable magnetic state remains the same with changing the position of the defective frameworks and their relative distance, which is beneficial for practical application.

(iii) GNR based p-n junctions

The non-covalent type doping is attractive because the fabrication is not as complicated as the electrostatic gate doping and it almost preserves the desirable electronic structure of GNRs unlike the chemical doping. Our group studied the current rectifying behavior of armchair-edge GNRs under a finite bias voltage [Carbon, 153, 525 (2019)]. It has been predicted that the rectification quality improves when the potential profile at the junction becomes abrupt, which is caused by the tunneling behavior. Indeed, our first-principles calculation results are consistent with the prediction.

3. Catalytic materials for electrochemical reactions

Clean energy is the worldwide issue nowadays. To reduce the pollutants, industries are now shifting towards using electricity rather than fossil fuels. In the energy conversion from chemical to electrical or vice versa, high efficiency of electrochemical reactions is one of essential criteria. The reaction rate is mainly governed by the intrinsic electrocatalytic activity where an ideal catalyst should bind the reaction intermediate neither too strongly nor too weakly. Also scarce elements should be avoided for commercial deployment. Thus finding a highly efficient and earth-abundant electrocatalyst is needed for sustainable future. Our group studies various types of electrocatalyst used for the oxygen reduction reaction (ORR), the oxygen evolution reaction (OER), hydrogen evolution reaction (HER), CO₂ and N₂ reduction reactions. It will help to design rationally the low-cost and innovative catalysts. In addition, from lesson of the structure-property relationship with mechanistic electronic structure analysis, one can apply to high-throughput machine-learning screening with a good descriptor.

4. Developing sophisticated simulation tools

First-principle calculations enable us to predict various properties of materials, such as atomic positions, electronic states, and resulting physical and chemical properties. Well-known density functional theory (DFT) has proven a powerful tool. It can describe rather an extended electronic system like solids and also a localized system like molecules with an ad-hoc term U, both of which are roughly categorized to metals and insulators, respectively. However the DFT-based method is not suitable to study strongly correlated materials which are neither metals nor insulators. Recently dynamical mean field theory (DMFT) has been a powerful tool. By combining the three methods, we can calculate or predict a desired property without relying on empirical parameters.

Sophisticated calculations of heterojunction electronic structure The density functional theory (DFT) method is very powerful to predict the physical and chemical properties of materials with relatively low computational cost. However, it suffers from the band gap underestimation due to the self-interaction error and poor description of the non-local screening. Although it can provide a meaningful trend over the material space, the quantitative variation of error is unpredictable, and it might not be helpful to reliable prediction of band alignment for the heterojunction. The state-of-the-art method is an empirical approach with the hybrid-type exchange-correlation functional, but its accuracy is not guaranteed for wide range of materials. The more reliable one has been the GW method with physically consistent description of carrier (or quasiparticle) injection/extraction. Given a weak interfacial coupling between A and B, the non-local screening can be described by the random phase approximation. Our group plans to study an efficient and accurate calculation method to predict the heterojunction band alignment. One technical process to be resolved is the Brillouin zone folding when there is the lattice mismatching.

Quantum transport in non-equilibrium heterojunction Charge and energy transport in a many-body quantum system at the nanoscale heterojunction exhibits many interesting phenomena due to strong couplings like electron-electron, electron-phonon. Most theoretical studies have been the model description, while recently real-time propagation and scattering of Kohn-Sham orbitals become popular. In order to describe the steady-state property, the non-equilibrium Green’s function approach (NEGF) is suitable. For proper description of the strong electron-electron correlation effect, our group have been studying to go beyond the non-crossing approximation for the impurity spectra in the framework of the pseudoparticle approach. We have obtained some results of Kondo phenomena for a single/double impurity orbital model. In future, our group plans to extend this to NEGF+DMFT within DFT. For this, we have implemented the DMFT in the Siesta package.