서울대학교 · Materials Science
한승우 교수의 연구실은 전자구조 계산과 분자역학 시뮬레이션을 기반으로 한 이차원 물질, 전이금속 디 chalcogenide, 나노소재의 물리화학적 성질을 깊이 있게 탐구하고 있습니다. 특히, 나노스케일에서의 결함, 표면, 나노튜브 등의 구조적 특성이 전기적, 화학적 성질에 미치는 영향을 이론적 방법으로 규명하며, 에너지 저장, 센서, 촉매 등 응용 가능성을 탐색하고 있습니다. 메시지 전달 기반 그래프 신경망을 활용한 원자간 잠재에너지 모델링을 통해 고정밀 시뮬레이션의 효율성도 동시에 추구하고 있습니다.
표시된 성과는 수집된 데이터 기준으로 산출되며, 일부 차이가 있을 수 있습니다.
Message-passing graph neural network interatomic potentials (GNN-IPs), particularly those with equivariant representations such as NequIP, are attracting significant attention due to their data efficiency and high accuracy. However, parallelizing GNN-IPs poses challenges because multiple message-passing layers complicate data communication within the spatial decomposition method, which is preferred by many molecular dynamics (MD) packages. In this article, we propose an efficient parallelization
Transition metal dichalcogenides (TMDs) have attracted enormous attention in diverse research fields. Especially, gas sensors are considered in a promising application exploiting TMDs. However, the studies are confined to only major TMDs such as MoS<sub>2</sub> and WS<sub>2</sub>. Particularly, the chemoresistive sensing properties of two-dimensional (2D) NbS<sub>2</sub> have never been explored. For the first time, we report room temperature NO<sub>2</sub> sensing characteristics of 2D NbS<sub>
We have performed $\mathrm{ab}$ $\mathrm{initio}$ pseudopotential electronic structure calculations for various edge geometries of the $(n,n)$ single-wall nanotube with or without applied fields. Among the systems studied, the one with a zigzag edge exposed by a slant cut is found to be the most favorable for emission due to the existence of unpaired dangling bond states around the Fermi level. The next most favorable geometry is the capped nanotube where $\ensuremath{\pi}$-bonding states locali
The catalytic activity for the hydrogen evolution reaction (HER) at the anion vacancy of 40 2D transition-metal dichalcogenides (TMDs) is investigated using the hydrogen adsorption free energy (Δ G<sub>H</sub>) as the activity descriptor. While vacancy-free basal planes are mostly inactive, anion vacancy makes the hydrogen bonding stronger than clean basal planes, promoting the HER performance of many TMDs. We find that ZrSe<sub>2</sub> and ZrTe<sub>2</sub> have similar Δ G<sub>H</sub> as Pt, th
We report an extensive ab initio study of self-interstitials in V and Mo. Contrary to the widely accepted picture, the $〈111〉$ dumbbell is found to be the most stable structure. The activated state for migration is the crowdion configuration, with an extremely low barrier $(\ensuremath{\sim}0.01\mathrm{eV}),$ suggesting $1d$ (one-dimensional) diffusion at low temperatures and $3d$ diffusion at high temperature. In the case of Mo, the energy landscape between the $〈111〉$ and $〈110〉$ dumbbells is
The native point defects in Fe2O3 are theoretically investigated using ab initio methods based on the GGA + U formalism. We consider vacancies and interstitials of Fe and O atoms as well as the electron polaron as Fe(II) defects at the host Fe(III) site. The formation energies and charge transition levels are computed for each defect type with careful elimination of size effects of the supercell. It is found that the Fe interstitial and vacancy form donor and acceptor levels close to band edges,
The ability to predict the behavior of point defects in metals, particularly interstitial defects, is central to accurate modeling of the microstructural evolution in environments with high radiation fluxes. Existing interatomic potentials of embedded atom method type predict disparate stable interstitial defect configurations in vanadium. This is not surprising since accurate first-principles interstitial data were not available when these potentials were fitted. In order to provide the input i
Two-dimensional (2D) molybdenum disulfide (MoS<sub>2</sub>) has been attracting rapidly increasing interest for application in chemoresistive gas sensors owing to its moderate band gap energy and high specific surface area.
Field emission properties of the (10,10) carbon nanotube are investigated with a first-principles approach. Emission currents are obtained through integrations of the time-dependent Schr\"odinger equation. We find that the emission current from the states localized at the tip end is more than ten times greater than direct contributions from extended metallic $(\ensuremath{\pi}$ and ${\ensuremath{\pi}}^{*})$ states. The spatial distribution of the electronic wave function as it tunnels through th
Throughout the past decades, doped-ZnO has been widely used in various optical, electrical, magnetic, and energy devices. While almost every element in the Periodic Table was doped in ZnO, the systematic computational study is still limited to a small number of dopants, which may hinder a firm understanding of experimental observations. In this report, we systematically calculate the single-element doping property of ZnO using first-principles calculations. We develop an automation code that ena
Combining total energy calculations with a search of phase space, we investigate the microscopic fusion mechanism of ${\mathrm{C}}_{60}$ fullerenes. We find that the $(2+2)$ cycloaddition reaction, a necessary precursor for fullerene fusion, may be accelerated inside a nanotube. Fusion occurs along the minimum energy path as a finite sequence of Stone-Wales transformations, determined by a graphical search program. Search of the phase space using the ``string method'' indicates that Stone-Wales
An efficient computational scheme based on the first-principles pseudopotential method is proposed for the electron emission from nanostructures under an applied electric field. The emission rate of the electron through the potential barrier is calculated by integrating the time-dependent Schr\"odinger equation for the states residing initially inside the emitter. Our approach takes into account the three-dimensional feature of the nanostructure as well as the realistic self-consistent potential
Intermolecular structures of porous two-dimensional supramolecular networks are studied using scanning tunnelling microscopy combined with density functional theory calculations. The local configurations of halogen bonds in polymorphic porous supramolecular networks are directly visualized in support of previous bulk crystal studies.