GIST Scholar
Not a member yet
30271 research outputs found
Sort by
On-chip terahertz emission from Floquet-Bloch states
No full textFloquet engineering uses time-periodic electromagnetic fields to modify the electronic properties of quantum materials via the creation of Floquet-Bloch states. These photon-dressed states inherit features from both the material and the driving field, enabling the exploration and control of quantum phenomena in light-matter hybrid systems. In non-centrosymmetric materials, shift currents can arise from the quantum geometric properties of electronic wavefunctions. However, shift currents from Floquet-Bloch states remain experimentally unexplored. Here, we employ an on-chip optoelectronic circuit to detect intrinsic terahertz emission from Floquet-Bloch states in Td-WTe2 under intense optical driving. We observe strong edge-localized terahertz emission that scales linearly with the driving field, consistent with the theoretical prediction for shift currents generated by Floquet-Bloch states. The results advance our understanding of strongly driven quantum materials and provide insights for developing efficient, bias-free terahertz sources for future optoelectronic technologies.TRUEsciescopu
ChatHAP: A Chat-Based Haptic System for Designing Vibrations through Conversation
No full textIn contrast to design tools for graphics and audio generation from text prompts, haptic design tools lag behind due to challenges in constructing large-scale, high-quality datasets including vibrations and text descriptions. To address this gap, we propose ChatHAP, a conversational haptic system for designing vibrations. ChatHAP integrates various haptic design approaches using a large language model, including generating vibrations using signal parameters, navigating through libraries, and modifying existing vibrations. To further improve vibration navigation, we present an algorithm that adaptively learns user preferences for vibration features. A user study with novices (n=20) demonstrated that ChatHAP can serve as a practical design tool, and the proposed algorithm significantly reduced task completion time (38%), prompt quantity (25%), and verbosity (36%). The study found ChatHAP easy-to-use and identified requirements for chat-based haptic design as well as features for further improvement. Finally, we present key findings with ChatHAP and discuss implications for future work. © 2025 Copyright held by the owner/author(s)
Discovery of catalysts for amidation and transesterification reactions using fluorescence-based high-throughput screening methods
No full textCatalyst development for organic reactions often requires selecting highly active catalysts from numerous candidates, but traditional evaluation of catalyst activity is time-consuming and labor-intensive. To overcome these vulnerabilities, high-throughput screening (HTS) methods using colorimetric or fluorometric responses have been developed, which enable rapid sample processing using relatively inexpensive equipment compared to conventional analytical techniques such as chromatography, mass spectrometry, and nuclear magnetic resonance. In this study, we developed fluorescence-based HTS methods utilizing fluorescent probes to enable rapid catalyst evaluation for organic reactions. Using the developed HTS methods, the most efficient catalysts for amidation and transesterification reactions were identified, and the applicability of these selected catalysts were subsequently verified.
Direct amidation of carboxylic acids with amines holds significant importance; therefore, catalytic processes involving boronic acids have undergone extensive investigation. However, studies focused on the amidation of aromatic carboxylic acids remain limited. Various boronic acid catalysts were evaluated for amidation using a new fluorescence-based HTS method that utilized the fluorescence turn-on characteristics of an anthracene-based probe in response to the amidation reaction. Our findings reveal that 2- hydroxyphenylboronic acid, previously deemed inefficient for aliphatic acids, effectively catalyzes the amidation of aromatic acids. The catalysts identified through this method consistently achieved high yields, reaching up to 98% across a broad spectrum of substrates.
The development of heterogeneous metal oxide catalysts for transesterification reactions is crucial owing to their seamless reusability and environmental friendliness. In recent years, numerous studies have been conducted on rare-earth oxides, such as lanthanide metal oxides. Various metal oxides were screened for transesterification using another new fluorescence-based HTS method that employed ratiometric fluorescence change of a pyrene excimer probe. Praseodymium (IV) oxide yielded the highest catalytic activity among the prepared metal oxides. Various substrates were successfully transesterified, and biodiesel was also produced in a high yield (90%) from soybean oil through transesterification using the catalyst. The selected catalyst required minimal amounts for the transesterification of various organic substrates (0.7 mol%) and soybean oil (0.8 wt%).DoctorAbstract i
Contents iii
List of Tables vii
List of Figures viii
I. Introduction to fluorescence-based high-throughput screening 1
1.1. High-throughput screening in catalyst development 1
1.2. Optical response-based high-throughput screening methods – Colorimetric 1
1.3. Optical response-based high-throughput screening methods – Fluorometric 5
1.4. Principles of fluorescence 9
1.5. Fluorescence modulation 10
1.6. Advantages of fluorescence-based analysis over colorimetric or absorbance-based
analysis 13
1.7. Research Overview 13
II. Discovery of boronic acid catalyst for direct amidation of aromatic carboxylic acids
using fluorescence-based high-throughput screening 14
2.1. Introduction 14
2.1.1. Diverse strategies to synthesis amides 14
2.1.2. Catalytic strategies to synthesis amides 16
2.1.3. Catalytic direct amidation reactions 18
2.1.4. Research overview 21
2.2. Results and discussion 22
2.2.1. Development of an HTS method for amidation with an anthracene-based
fluorescent probe 22
2.2.2. HTS of boronic acid derivative catalysts 26
2.2.3. Optimization of reaction conditions 29
2.2.4. Substrate scope 31
2.2.5. Putative mechanism for direct aromatic amidation with catalyst C7 33
2.3. Conclusions 34
2.4. Experimental 35
2.4.1. Materials and instruments 35
2.4.2. Synthesis and characterization of fluorescent probe and amidation product 35
2.4.3. Procedure for fluorescence-based HTS of amidation 37
2.4.4. General procedure for amidation of aromatic carboxylic acids with amines 37
2.4.5. Characterization of amide products 38
III. Discovery of lanthanide metal oxide catalyst for transesterification reaction using
fluorescence-based high-throughput screening and application to biodiesel production 43
3.1. Introduction 43
3.1.1. Various catalytic strategies for transesterification 43
3.1.2. Metal oxide catalysts for transesterification 46
3.1.3. Biodiesel production by transesterification 47
3.1.4. Rare-earth oxide catalyst for transesterification 48
3.1.5. Research overview 49
3.2. Results and discussion 50
3.2.1. Development of HTS method for transesterification with pyrene excimer
fluorescent probe 50
3.2.2. HTS of transition or lanthanide metal oxide catalysts 54
3.2.3. Optimization of reaction conditions 58
3.2.4. Substrate scope 60
3.2.5. Application to biodiesel production 63
3.3. Conclusions 65
3.4. Experimental 66
3.4.1. Materials and instruments 66
3.4.2. General procedure for preparation of transition or lanthanide metal oxides 66
3.4.3. Synthesis of pyrene excimer probe, bis(4-(1-pyrenyl)butyl) maleate (BPBM) 66
3.4.4. General procedure for fluorescence-based HTS of transesterification 67
3.4.5. General procedure for transesterification of aryl esters with alcohols 68
3.4.6. Transesterification of glyceryl trioctanoate 68
3.4.7. Production of biodiesel by transesterification of soybean oil 68
3.4.8. Characterization of ester products 69
References 72
Appendix S 1
List of spectral copies of 1H, 13C, 19F NMR S 1
Spectral copies of 1H, 13C, 19F NMR S 5
Curriculum Vita
Reinforcement learning for the design of targeted antimicrobial peptides against resistant pathogens
No full textAntimicrobial peptides (AMPs) are considered the most promising alternative to traditional antibiotics, addressing the issue of antibiotic resistance through diverse mechanisms that differ from conventional antibiotics. However, failures in clinical trials of AMPs highlight the importance of balancing low toxicity with high activity.
This study proposes a novel AMP generation model that integrates reinforcement learning, Generative Pretrained Transformer (GPT), and Low-Rank Adaptation (LoRA) to design peptides that account for both activity and non-hemolytic properties for each strain. The goal of this thesis is to contribute to the development of potential AMPs by ensuring activity and simultaneously generating non-hemolytic peptides tailored to individual strain.
To validate the strain-aware generation, the embeddings of peptides generated for each strain were visualized using t-SNE plots. Furthermore, the model demonstrated the ability to generate peptides satisfying both activity and non-hemolytic conditions for unseen strains, highlighting its potential to design AMPs for emerging pathogens.MasterAbstract i
Contents ii
List of tables iv
List of figures v
I. Introduction 1
1. 1. Antibiotic resistance 1
1. 2. Antimicrobial peptides (AMPs) 2
1. 3. Existing methods for AMP generation 3
1. 4. Research Objectives 5
II. Materials and Methods 7
2. 1. Datasets and feature extraction 7
2. 2. MIC regression model 9
2. 3. Hemolysis prediction model 10
2. 4. Causal language modeling for pretraining 11
2. 5. Reinforcement learning 12
2. 6. Hyper-parameter tuning 14
2. 7. Evaluation metrics and prediction tools 15
III. Results and Discussion 16
3. 1. Pretraining performance 16
3. 2. Diversity, uniqueness, novelty 18
3. 3. Hemolysis prediction performance 19
3. 4. MIC regression performance 21
3. 5. MIC & Hemolysis performance 22
3. 6. T-SNE plot of generated peptides 23
3. 7. Amino acid composition 24
3. 8. Top 3 structure predictions 25
3. 9. Unseen strain generation performance 26
IV. Conclusion 28
Summary 29
References 30
Acknowledgement 34
Curriculum Vitae 3
Adsorptive Removal of Per- and Polyfluoroalkyl Substances (PFAS) by Polyaniline (PANI) and Chitosan-graft-polyaniline (CS-PANI)
No full textProton Accumulation Modulated Surface Potential in Proton Conducting Ceramics Revealed by Near-Ambient Pressure X-ray Photoelectron Spectroscopy
No full textProton conducting electrochemical cells (PCECs) are efficient and clean intermediate-temperature energy conversion devices. The proton concentration across the PCECs is often nonuniform, and characterizing the distribution of proton concentration can help to locate the position of rate-limiting reactions. However, the determination of the local proton concentration under operating conditions remains challenging. Here, we employed in situ near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) to investigate an Au/BaZr0.9Y0.1O3-δ/Au symmetric cell with DC bias of 1 V applied between the working and counter electrodes (CE). The relative intensity of hydroxyl groups, deconvoluted from the O 1s XPS spectra, reveals the distribution of proton concentration across the electrolyte. The applied electric field induces proton accumulation at the counter electrode, imposing binding energy shifts of the surface components for metal elements relative to their lattice components. Combined XPS and impedance analysis suggests that the accumulation layer of protons is much thicker at 500 K compared to that at 670 K, as a result of a larger amount of hydroxyl groups at the lower temperature. This nonuniform distribution of proton concentration affects the chemical environment of metal elements, and the local electrical potential, as revealed by the in situ XPS. This work demonstrates in situ NAP-XPS as a tool to probe the distribution of proton concentration and its impact on the defect chemistry and local electrical potential of PCECs, thereby advancing the understanding of the impact of proton defect chemistry and the performance improvement of PCECs. © 2025 American Chemical Society.FALSEsciescopu
Intelligent Metallic Loose Part Monitoring in Steam Generator Using Convolutional Neural Networks and the Position-Invariant Loss Function
No full textThe degradation and aging of primary system in nuclear power plants, as well as the potential ingress of Loose parts during planned maintenance, can result in internal impacts caused by loose parts within the primary system. When these loose parts collide with the welded tube sections in the lower part of the steam generator, cracks may develop, potentially leading to coolant leakage from the primary side to the secondary system. To prevent such incidents, nuclear power plants have implemented Loose Parts Monitoring Systems (LPMS). Current LPMS employ rule-based algorithms to estimate the impact location and mass of loose parts. However, this approach has notable drawbacks, including variability in analysis results depending on the proficiency of signal analysts and the considerable time required for signal analysis. To address these limitations, this study proposes a Convolutional Neural Network (CNN)-based model for estimating the mass and impact location of loose parts. The proposed model incorporates a Position-invariant Loss Function based on Cartesian coordinates and uses the WignerVille distribution as input. The proposed method was validated using impact signals obtained from a 1/4-scale steam generator testbed designed to simulate a primary system steam generator. Experimental results demonstrate that the proposed approach can estimate the impact location and mass of foreign objects in three-dimensional structures more quickly and accurately than conventional methods.|합성곱 신경망과 위치 불변 손실함수를 이용한 증기발생기의 지능형 금속파편 감시 시스템 원자력 발전소 일차계통 구조물 열화 및 노후화로 인한 이물질 발생 또는 계획 정비 기간 동안 외부 이물질 유입 등 다양한 이유로 일차계통 내부에 이물질에 의한 내부 충격이 발생할 수 있다. 내부 이물질이 증기발생기 하부 세관 용접 부위에 충돌하게 되면 균열이 발생하고, 이 부위를 통해 냉각수가 2차측으로 유입 되는 사고가 발생할 수 있다. 이러한 사고를 예방하기 위해 원전에서는 LPMS(Loose Parts Monitoring System)를 설치하여 운영하고 있다. 기존 LPMS는 Rule base 알고리즘을 적용하여 내부 이물질 충돌위치 및 질량을 추정하고 있다. 따라서, 알람 신호에 대해 신호분석 전문가의 능숙도에 따라 분석결과가 달라지고, 신호분석에 시간이 많이 걸린다는 단점이 있다. 이러한 문제점을 해결하기 위해 본 논문에서는 신속하고 정확한 분석을 위해 데카르트 좌표계에 기반한 Position-invariant Loss function과 위그너-빌 분포를 입력으로 하는 Convolutional Neural Network 기반 이물질 질량, 충격 위치 추정 모델을 제안한다. 1차 계통 증기발생기를 모사하는 1/4 스케일 테스트베드에서 취득된 충격 신호를 사용하여 제안된 방법을 검증한다. 본 연구에서 제안된 방법이 3차원 구조물에서 내부 이물질의 충격 위치와 질량을 기존 방법에 비해 더 빠르고 정확하게 추정할 수 있음을 실험적으로 확인한다.MasterAbstract i
Contents iii
List of Figures iv
List of Tables vi
Chapter 1. Introduction 7
Chapter 2. Steam Generator in the Primary System 10
Chapter 3. Proposed Methodology 12
3.1. Impact Signal Pre-processing 12
3.2. Proposed Neural Network Architecture 15
3.3. Position-invariant Loss function 16
3.4. Comparison Strategies 19
Chapter 4. Results and Discussion 22
4.1. Experimental setting 22
4.2. Quantitative Performance Evaluation 24
4.3. Qualitative Performance Evaluation 27
4.4. Training parameters 34
4.5. The Possibility of Online Monitoring 35
Chapter 5. Conclusions 37
References 39
Curriculum Vitae 4
Data-driven Precipitation Nowcasting Using Satellite Imagery
No full textAccurate precipitation forecasting is crucial for early warnings of disasters, such as floods and landslides. Traditional forecasts rely on ground-based radar systems, which are space-constrained and have high maintenance costs. Consequently, most developing countries depend on a global numerical model with low resolution, instead of operating their own radar systems. To mitigate this gap, we propose the Neural Precipitation Model (NPM), which uses global-scale geostationary satellite imagery. NPM predicts precipitation for up to six hours, with an update every hour. We take three key channels to discriminate rain clouds as input: infrared radiation (at a wavelength of 10.5 mu m), upper- (6.3 mu m), and lower-(7.3 mu m) level water vapor channels. Additionally, NPM introduces positional encoders to capture seasonal and temporal patterns, accounting for variations in precipitation. Our experimental results demonstrate that NPM can predict rainfall in real-time with a resolution of 2 km
High-fidelity simulation of laminar-to-turbulent transition in hypersonic boundary layer on a sharp cone
No full textLaminar-to-turbulent transition in hypersonic boundary layer on a straight cone is numerically investigated in this study. High-fidelity simulation is performed with direct-numerical simulation (DNS) coupled with the linear stability theory (LST). This study focuses on the transition scenario of fundamental breakdown, driven by the two-dimensional Mack 2nd mode and a pair of oblique modes. The major instabilities in the hypersonic boundary layer are identified by LST and introduced at the DNS inlet. Current DNS computations successfully capture intrinsic transition phenomena, including aligned vortical structures and peak heat flux in the transition process, and complete transition to turbulent flow. Appropriate numerical dissipation associated with shock-capturing methods is investigated in this study because of the presence of a nose shock outside the boundary layer and compression waves from amplified instabilities inside the boundary layer. This numerical study is conducted with two shock sensors. A classical shock sensor generates excessive dissipation in the viscous boundary layer, which artificially delays the turbulent transition. The alternative sensor reduces the unintended dissipation, allowing flow to develop turbulence within the computational domain. Computational data are discussed with relevant experimental and theoretical data. © 2025 Author(s).FALSEsciescopu