이 논문에서는 제일원리계산을 수행하여 VSe2 단일층의 전자 구조를 연구하였다. VSe2 단일층의 이론 격자 상수는 3.33Å, 단일층의 높이는 3.16Å로 계산되었다. VSe2 단일층은 0.60μB의 자기모멘트를 지니며, 이러한 자성에는 V 3d orbital이 주로 기여한다. 에너지 밴드 구조는 페르미 준위 근처에서 spin-up 밴드와 spin-down 밴드의 spin splitting이 0.43eV 정도로 나타나는 것을 확인하였다. 최종적으로 계산된 밴드 구조를 이용하여 Fermi surface를 구하고, Fermi surface nesting이 존재함을 확인하였다. 이는 1T-VSe2 단일층이 전하밀도파를 형성할 가능성이 높다는 것을 의미한다.

It is recently reported that thin-layered SnSe2 can convert CO2 into formate with high efficiency. In this study, mechanisms for CO2 reduction reaction (CO2RR) and hydrogen evolution reaction (HER) on SnSe2 monolayer are studied by density functional theory (DFT) calculations. The basal plane and edge were not possible active sites, different from most dichalcogenides having edge as an active site. The defects on SnSe2, especially Se vacancies, were responsible for the good catalytic properties of the monolayer, not the metallic Sn formed by oxidation of SnSe2. Suggesting the active sites, this study can be used for catalyst design in the future.

이 논문에서는 이차원 층상 구조 물질인 TaS2의 2H와 1T polytype에 대한 제일원리 전자구조계산을 수행하였다. 그 결과, 2H 구조가 1T 구조보다 총 에너지가 0.07eV/unit cell 낮아 더 안정함을 확인하였다. 에너지 밴드 구조와 전자상태밀도 계산을 통해 두 구조 모두 금속성을 띰을 확인하였다. 계산된 밴드 구조로부터 얻어지는 Fermi surface는 2H 구조에서 hole pocket만 형성되고, 1T 구조에서 electron pocket만 형성된다. 1T 구조의 Fermi surface는 nesting으로 인해 전하밀도파 현상이 일어날 가능성이 높음을 암시한다.

Graphene은 2004년 Geim, Novoselov에 의해 최초로 분리된 이후, zero bandgap의 dirac cone 모양의 band structure를 가지는 전기적 특성에 의해 주목받고 있는 물질이다. 하지만 Geim과 Novoslev의 분리법은 대량 생산에 문제점이 있어 새로운 공정들이 많이 개발되어 왔다. 특히, graphite를 산화시키고 이를 다시 환원시켜 단층의 reduced graphene oxide를 만드는 공정이 현재 각광받고 있다. 하지만 이 공법은 oxidation과 reduction을 반복하면서 oxygen과 carbon이 결합하는 defect가 생기는 단점이 존재한다. 이번 연구에서는 reduced graphene oxide에서 특징적으로 발생되는 Epoxy functional group 및 Hydroxy functional group defect가 graphene의 band structure에 어떠한 영향을 끼치는지 분석하였다. 연구 결과, 이 functional group defect들은 graphene의 zero band gap을 open시키는게 확인되었고, Epoxy group은 p-dopant로 작용하지 않는 반면 Hydroxy group은 p-dopant로 작용함을 확인하였다. 이러한 차이점은 두 group과 결합한 C 원자의 relaxation structure를 확인해 볼 때 다른 hybrid orbital을 가지기 때문으로 해석할 수 있다.

Ultra thin body Tunnel field-effect transistors (UTB TFETs) 의 channel 을 Si 를 사용하여 biaxial strain engineering 에 의한 소자의 성능을 비교, 분석하였다. 본 연구에서는 device 가 scaling 됨에 따라 매우 얇은 thickness 를 선택하고자 하였고, band gap 이 너무 커지지 않는 선에서 2 nm 를 시뮬레이션에 사용하였다. 먼저 compressive strain 을 인가했을 때 3% 까지는 소자의 성능 변화가 두드러지지 않았다. 반면에, Tensile strain 을 인가했을 때는 3% 일 때 no strain (0%) 과 비교하여 on current 가 5 배 가까이 증가함을 확인할 수 있었다. 기존에 연구되었던 compressive, tensile strain 외에도 asymmetric strain 에서는 3% 일 때 off current 가 증가함으로써 TFETs 본래의 장점인 low power application 으로써의 활용이 어려움을 파악하였다. 따라서 본 연구에서는 asymmetric strain effect 가 소자의 성능에 critical factor 로 작용할 수 있다는 것을 확인하였다.

The alltrope of carbon has been steadily studied because it shows unique physical properties. Since fullerens were observed for the first time in 1985 [1], carbon system has attracted attention in the field of nanoelectronics in that the remarkable properties could open many opportunities beyond Moore’s law. Among the allropes, graphene, which consists of sp2-hybridized carbon network system with a linear energy dispersion relation, has been the most intensely studied because of its high electrical, thermal mobilities. However, the applications to nanoelectronics are hindered because pristine graphene has zero band gap which limits achievable on-off current ratios.[2] As an alternative materials, other alltropes with non-zero band gap, such as graphane, graphene oxide and graphyne, has attracted much attention. Of that, graphyne especially has interesting electronic properties. It was theoretically found that band gap of graphyne can be linearly tunned by strain.[3] Graphyne can have advantages over graphene in nanodevice design and application due to the property of band gap variation under strain.

본 연구에서는 3차원 집적 회로에 적용되는 silicon-on-insulator(SOI) field-effect transistor(FET)의 채널 도핑의 유형 및 농도에 따른 DC 성능 변화 및 열화 현상에 대해 조사했다. EDISON에서 제공하는 “UTB FET” 시뮬레이터를 활용하여, 7 나노 노드의 SOI FETs의 채널의 두께, 채널 도핑의 유형 및 농도, 소자 온도에 따른 전류-전압 특성을 얻고, 이를 에너지 밴드 다이어그램을 통해 분석했다. 그 결과, 채널의 두께가 3 nm로 작을 때 채널 도핑의 유형 및 농도에 상관없이 단채널 효과의 감소로 소자의 구동 전류가 증가한다. 한편, 소스/드레인의 도핑 농도는 작지만 채널 도핑 농도와 동일한 junctionless FET는 유효 채널 길이의 증가로 단채널 효과가 더욱 작아지며, 저전력 응용분야에서 구동 전류가 크게 증가한다. 또한 다른 SOI FETs와는 달리 동작 지점에서 에너지 장벽이 없어 온도 변화에 따른 구동 전류 변화가 작다. 본 논문의 성능 및 열화 분석을 통해, junctionless FET가 향후 5 나노 노드 및 3차원 집적 회로에 적용될 것으로 기대된다.

We have investigated the performances of MXenes as catalysts and transistors. To evaluate the excellence of MXene-based-catalyst, free energies and exchange currents were compared. We found Ti3C2O2 is the best MXenes for HER, followed by Ti4C3O2. MXene transistors are modelled by DFT and transport simulation was proceeded by NEGF-Poisson self-consistent simulation in the mode space, under ballistic condition. MXene transistors were less sensitive to the scaling down of the devices, when compared by BP. Ti2CO2, Zr2CO2 and Hf2CO2 AD-6nm-nFETs show sub-threshold swing of 60 mV/dec while AD-6nm-BP exhibits 261 mV/dec because they have higher effective masses of electrons, meaning less direct-tunneling occurs under off-region. Among of them, Both 10-nm Zr2CO2 nFET and pFET exhibit isotropic transport with the on-state current over 2350 μA/μm, Channel capacitance over 7 μF/cm2 and swing under 68mV/dec, meeting all requirements of 2021 of ITRS-2017. We strongly suggest Zr2CO2 FET is the one of the promising candidate of future logic devices.

금속-절연막-금속의 패턴을 가지는 소자 특성 분석 시뮬레이션을 진행하였다. 전극은 백금, 절연막으로는 보편적으로 RRAM 소자 연구에 사용되는 HfO2와 TiO2, Ta2O5를 사용했으며 절연막의 두께, 트랩의 수, 트랩의 깊이 등을 변수로 하여 각 RRAM 소자의 전류 특성을 비교하였다. 트랩의 수, 트랩의 깊이를 변화시켰을 때, 각각의 물질 내에서는 RRAM을 작동시키기 위한 Transition Voltage와 Transition Field가 두께에 따라 차이가 거의 없었다. 하지만, 절연막의 두께가 얇아질수록, Transition Voltage는 작아지고 전류 밀도는 커지므로 결과적으로, 저항이 감소함을 알 수가 있었다. 절연막으로써 사용되는 세 물질을 비교했을 때, Ta2O5가 다른 물질보다 더 좋은 특성을 가짐을 확인 할 수 있었다.

CdSe quantum dot is a promising route for realization of low-cost, large-area, flexible and light weight photovoltaic devices. However, its stability and electron-phonon coupling that lower the efficiency of device are pointed out as critical defect of CdSe quantum dot. Attaching ligands around quantum dot is suggested as solution by changing distribution of charge around the surface. By attaching hydrogen and 1,6-hexanedithiol (HDT), CdSe quantum dot’s electronic structure changes. The proper electronic structure is that H is terminated with 1.5e and 0.5e charges and HDT ligand substitutes H position to generate hole trapping energy level around valence band edge.

오픈 소스 반도체 소자 시뮬레이터인 G-Device를 소개합니다. G-Device는 C++ 언어로 개발되었으며, 개발자들이 구조를 쉽게 이해하여 자신의 문제에 적용할 수 있는 코드를 목표로 하고 있습니다. THz 주파수 대역에서의 drift-diffusion 시뮬레이션, Boltzmann 수송 방정식의 해석, 양자 전송 현상의 해석 등, 반도체 소자 시뮬레이션과 관련된 예제들을 설명합니다.

The Atomic Simulation Environment (ASE) package written in Python is aimed at providing an easy, flexible and customizable environment for density-functional-theory (DFT) calculations[1]. Currently, the ASE package supports interfaces over 30 Calculators including ABINIT, SIESTA, Quantum Espresso, and VASP. Currently, the ASE package supports interfaces over 30 Calculators including ABINIT, SIESTA, Quantum Espresso and VASP. While most of ab initio calculations require massive computational resources and opt for high efficiency, the Python language behind the ASE package provides a copious number of library modules which make swift code developments possible. Here, we like to take advantage of both the OpenMX code and the ASE/Python modules by implementing the OpenMX interface for the ASE. For example, combining the LAMMPS and OpenMX codes, we can seamlessly perform molecular dynamics simulations without compensating coding efficiencies. Even benchmarking the OpenMX results against VASP is a breeze. Since the modules in ASE are highly scalable, we can utilize some of the features of ASE modules, e.g., the phonon module, for the OpenMX calculations. We hope that this platform can be used for combining DFT calculations with machine learning and big data, where the famous Python modules like Spark and TensorFlow can be employed to conduct deep learning research.

본 연구에서는 SIESTA와 NEGF기반 양자트랜지스터 소스코드의 웹입출력 양식을 개발하였다. SIESTA 4.1.3b 버전을 이용하였다. 구현을 위해서 Javascript와 WebGL을 이용하였으며, Liferay portlet형태로 서비스되고 있다. 상호작용 가능한 3차원 인터페이스를 제작하였다.

The conventional effective mass approximation (EMA) could be an efficient method to treat nanostructures with sizes too large to be handled in first-principles calculations, but the employment of bulk effective mass and dielectric constant can be problematic to accurately describe them. Herein, we report on the development of a first-principles data-based EMA approach for the simulation of optoelectronic properties of finite nanostructures and its application to CdS/ZnS core/shell quantum nanorods, where relevant experimental data became recently available. We first carry out density functional theory (DFT) calculations for an infinite nanowire to obtain dielectric constant, hole and electron wavefunctions, and corresponding effective masses. The information is then fed into the finite nanorod cases to construct EMA wavefunctions, which are subsequently used to estimate their photoluminescence and field-dependent switching properties. Simulation results show that in contrast to the bulk parameters, the first principles-based EMA parameters are flexible enough to describe nanorods with different diameter and length. Furthermore, we find that exciton transition energies are very sensitive to the physical dimension of nanorods. We found that photo luminance (PL) quenching efficiency of different nanorods emitting same light color depends sensitively on their diameters. In this study, our proposed technique about modified mass and dielectric parameters within EMA, opens a unique way to explore quantum nanorod and nanoplatelet shape structures ranging from very small to very large sizes.

Recently, B₂O₃ has drawn attention for crystallization anomaly. One of oldest known crystal, alpha structure, emerges only in high pressure. Sigma bonding, beta structure, appears when extremely high pressure is applied[1] Although, in ambient pressure, it show energetically stable in crystal form, it resides as vitrified structure[2]. Here we calculated electronic structure of B₂O₃ using Edison app, and analyzed its physical properties.

One of the crucial bottlenecks of zigzag graphene nanoribbons (ZGNRs) in nanoelectronics is metallic properties, originated from gapless band structure, and thereby the normal ZGNRs cannot be utilized as a field effect transistors (FETs). However, ZGNRs can have bandgap when the proper boundary conditions with defective edge structures are formed. To delve into the semiconductor property of ZGNRs, we conducted the first principles based density functional theory (DFT) calculations and performed transport calculations based on the non-equilibrium Greens function (NEGF) formalism. We found that the bandgap can be enlarged to the extent of the 1 eV, which is prominently the semiconductor property, by the modulation of defective edge structures. Our overall results provide the conceptual motivation to delve into the importance of boundary conditions of ZGNRs. For the DFT calculations, DFT based quantum transistor simulator 1.0.0 simulator is used.

While one of the most anticipated initial applications of redox-controllable mechanically interlocked supramolecular complexes was as an active component in memory devices, some skepticisms were immediately raised on the molecular origin of observed switching behavior and after more than a decade from the initial report further progress still remains minimal. On the occasion of the 2016 Nobel Prize in Chemistry being awarded on “molecular machines”, we revisit and expand our earlier multiscale computational study on the switching function in supramolecular electronic devices [Phys. Rev. Lett. 94, 156801 (2005)]. Adopting the rotaxane case, we first find that the strong identity of supramolecular cores leads to a robust switching behavior. In spite of the different spatial arrangements of redox-active components, the electrical switching signals are similar to those observed from its [2]catenane counterpart, signifying the advantage of supramolecular complexes for molecular electronics. However, we find that even a single metal atom penetrated into the vicinity of these molecular system can destroy the electronic signature of molecular cores and thus the associated molecule-originated switching. In this case, one will instead observe a memristive switching originating from conducting nanoflaments. This work reconciles the contradicting interpretations on the origins of the electrical switching observed in supramolecular electronic devices and paves the way for further developments of nanoelectronic devices based on molecular machines.