GIST Scholar
Not a member yet
30271 research outputs found
Sort by
In situ magnetic-field-assisted bioprinting process using magnetorheological bioink to obtain engineered muscle constructs
No full textTissue-engineered anisotropic cell constructs are promising candidates for treating volumetric muscle loss (VML). However, achieving successful cell alignment within macroscale 3D cell constructs for skeletal muscle tissue regeneration remains challenging, owing to difficulties in controlling cell arrangement within a low-viscosity hydrogel. Herein, we propose the concept of a magnetorheological bioink to manipulate the cellular arrangement within a low-viscosity hydrogel. This bioink consisted of gelatin methacrylate (GelMA), iron oxide nanoparticles, and human adipose stem cells (hASCs). The cell arrangement is regulated by the responsiveness of iron oxide nanoparticles to external magnetic fields. A bioprinting process using ring magnets was developed for in situ bioprinting, resulting in well-aligned 3D cell structures and enhanced mechanotransduction effects on hASCs. In vitro analyses revealed upregulation of cellular activities, including myogenic-related gene expression, in hASCs. When implanted into a VML mouse model, the bioconstructs improved muscle functionality and regeneration, validating the effectiveness of the proposed approach. © 2024 The AuthorsTRUEsciescopu
The release of chemical additives from simulated abrasion tire wear particles: Insights from ICP-MS and GC-MS analysis
No full textTire wear particles (TWPs), a major source of microplastic pollution, contain a complex mixture of chemical additives and are generated through the mechanical abrasion of tires. This study evaluates the physical characteristics and chemical toxicity of TWPs derived from four tires of varying mechanical strengths (T250, T350, T500, and T700). TWPs were produced in a controlled setting and characterized using microscopy, SEM, and FT-IR. While all samples shared a base composition of styrene-butadiene rubber, particle size and morphology differed with tire strength, with stronger tires generating larger, more irregular fragments. Leachate analyses using ICP-MS and GC-MS revealed the presence of both inorganic and organic contaminants. Zinc was the dominant metal, with concentrations reaching up to 4117 µg/L. GC-MS screening identified various organic compounds including benzothiazole, aniline, 6PPD, and BHT – known for their persistence and toxicity. Hazard classification based on GHS and UNEP criteria showed that many of the leached compounds are PBT or vPvB substances, posing significant ecological risks. The findings highlight that TWPs not only vary in physical form depending on tire strength but also release toxic additives capable of long-term environmental harm. This study emphasizes the importance of regulating chemical additives in tires and supports a more comprehensive approach to microplastic pollution assessment.MasterI. Introduction １
1.1 Tire wear particle is the new microplastics thread to the environment １
1.2 Chemical additive in TWPs １
1.3 Non-targeted chemical analysis １
1.4 Toxicity assessment ２
1.5 The objective of the study ２
II. Materials and Methods ３
2.1 Sample preparation and leaching experiment ３
2.2 Physical characterization ３
2.3 Chemical analysis ４
III. Results and Discussion ８
3.1 Characterisation of TWPs ８
3.2 Leachate compound identification ９
3.3 Toxicity assessment base on GHS hazard coding １８
IV. Conclusion ２５
V. References ２６
VI. Acknowledgements ３
Enhancing Zinc-Bromine Battery Durability with Biphasic Electrolytes
No full textAqueous static zinc-bromine batteries (ZBB) offer high theoretical energy density and low cost due to their simplified design compared to flow batteries, making them promising for energy storage systems (ESS). However, ZBBs face self-discharge issues due to bromine (Br2) and polybromide anions (Br3 crossover from the positive electrode, resulting in low coulombic efficiency (CE) and capacity fading. Also, the thermodynamic instability of zinc metal anode in aqueous solutions leads to challenges such as hydrogen evolution reaction (HER) and zinc dendrite growth. This study proposes a biphasic electrolyte system comprising aqueous and organic solutions to address challenges at both the cathode and anode. On the cathode side, an aqueous solution containing 0.1 M tetrapropylammonium bromide (TPABr), 0.5 M ZnBr2, and 1 M Zn(CH3COO)2 was used. On the anode side, diethyl carbonate (DEC) was chosen as the organic solvent due to its immiscibility with water and compatibility with zinc ions for effective stripping/plating. The high ionic conductivity of aqueous phase facilitates bromine redox reactions, while bromine crossover was suppressed by TPABr, retaining 86% capacity after a 48-h open-circuit period. This system suppressed parasitic reactions and zinc dendrite growth, resulting in high CE and extended cycle life, with more than 1,160 cycles at approximately 100% CE. In previously reported biphasic electrolytes, zinc reactions at the anode occurred in aqueous systems while bromine reactions at the cathode occurred in organic solvents, leading to zinc metal instability. Our approach overcomes these dual challenges by reversing the organic and aqueous phases. This work paves the way for future advancements in static battery design, potentially leading to more sustainable ESS.MasterAbstract
Contents
List of tables
List of Figures
Ⅰ. Introduction
Ⅱ. Experimental Section
Ⅱ.1. Materials
Ⅱ.2. Electrolyte Preparation
Ⅱ.3. Electrode Preparation
Ⅱ.4. Cell (Beaker cell & Zn-Br stack)
Ⅱ.5. Materials Characterization
Ⅱ.6. Electrochemical Measurements
Ⅲ. Result and discussion
Ⅲ.1. Optimizing Electrolyte Composition
Ⅲ.2. Physicochemical analysis of electrolyte properties
Ⅲ.3. Suppressing Br3− cross-diffusion at the cathode
Ⅲ.4. Characterization of zinc deposition behavior and dendrite suppression in BE
Ⅲ.5 Expansion of electrochemical stability window (ESW) due to the organic layer in BE
Ⅲ.6 Rate performance and long-term stability evaluation of BE
Ⅳ. Summary
Ⅴ. Reference
Acknowledgement
Curriculum Vita
Probing spatiotemporally organized GPCR signaling using genetically encoded molecular tools
No full textG-protein-coupled receptors (GPCRs) control various downstream signaling pathways, with multiple effectors whose interactions are subject to sophisticated regulation to achieve signaling specificity. Spatiotemporal organization of GPCR signaling is essential for efficient control of multifaceted signaling pathways. To study how this spatiotemporal signaling is structured and affects cellular functionality, various genetically encoded molecular tools that can detect and perturb the target biochemical activities at a subcellular level have been developed. In this Review, we introduce various types of fluorescent protein-based biosensors and molecular tools that allow us to directly elucidate the spatiotemporal mechanisms of GPCR signaling regulation at a subcellular level. Finally, we highlight several applications of these molecular tools to study the spatiotemporal organization of GPCR in living cells to obtain a comprehensive understanding of the signaling architecture.TRUEsciescopuskc
Enhancing RAG Performance with GA-based Weighted Reranker Optimization
No full textThis study presents a novel approach for optimizing reranker performance in Retrieval-Augmented Generation (RAG) systems using Genetic Algorithms (GA). Existing RAG systems typically rely on fixed weights or single reranker models, which fail to effectively capture the diverse characteristics of multiple reranker models. To address this limitation, we propose using GA to dynamically optimize the weights of reranker models by assigning optimal weights to each reranker and maximizing their characteristic that can enhance the accuracy of the final document ranking. Experiments conducted with 708 document-query pairs show that the GA-optimized reranker achieved an accuracy rate of 76.69%, significantly outperforming the highest-performing individual reranker, FlashRank (72.88%). This study demonstrates that GA optimization can significantly improve reranking performance in RAG systems and highlights its potential for application in complex domains such as finance. Additionally, the proposed GA-based optimization method effectively combines multiple reranker models to maximize document retrieval and response performance.|본 연구는 유전자 알고리즘(Genetic Algorithm, GA)을 활용하여 Retrieval-Augmented Generation (RAG) 시스템의 리랭커(Reranker)의 성능을 최적화하는 접근 방안을 제시한다. 기존의 RAG 시스템은 주로 고정된 가중치나 단일 리랭커 모델에 의존하여 여러 리랭커 모델의 다양한 특성을 충분히 반영하지 못한다는 한계가 있다. 이를 극복하기 위해 본 연구에서는 GA를 활용하여 각 리랭커 모델의 가중치를 동적으로 최적화하고, 이를 통해 리랭킹 성능을 개선해 최종 문서 순위 결정의 정확도를 높일수 있고자 하였다. 실험은 708개의 문서-질의 쌍을 사용하여 수행되었으며, GA 최적화 리랭커는 76.69\%의 정답률을 기록하며, 개별 리랭커 중 가장 높은 성능을 보인 FlashRank(72.88\%)를 초과하는 성과를 달성하였다. GA는 각 리랭커에 적합한 가중치를 할당하여 리랭커들의 장점을 극대화하며, 최종 문서 순위 결정의 정확도를 크게 향상시켰다. 본 연구는 GA 기반 최적화가 RAG 시스템의 리랭킹 성능을 크게 향상시킬 수 있음을 보여주며, 금융과 같은 복잡한 도메인에 적용할 잠재력을 지닌다는 점을 강조한다. 또한, 제안된 GA 기반의 최적화 방법이 다양한 리랭커 모델을 효과적으로 결합하여 문서 검색 및 응답 성능을 극대화하는 역할을 할 수 있음을 보여준다.MasterAbstract (English) i
Abstract (Korean) iii
List of Contents v
List of Tables vii
List of Figures viii
List of Algorithms ix
1 Introduction 1
2 Background and Theory 5
2.1 Overview of Retrieval-Augumented Generation Systems 5
2.2 Rerankers in RAG Systems 7
2.3 Metaheuristic Optimization and Genetic Algorithm 9
2.4 Optimization of Reranker Weights 11
3 Experiments 15
3.1 Datasets 16
3.2 Composition of a RAG System 18
3.2.1 Load 18
3.2.2 Split 18
3.2.3 Embeddings 19
3.2.4 Vector Database 19
3.2.5 Retrieval 19
3.2.6 Reranking 20
3.2.7 Prompting 25
3.2.8 LLM & Output 25
3.3 Genetic Algorithm Configuration 27
3.3.1 Document Combined Score 27
3.3.2 Components of Genetic Algorithm (GA) 29
3.4 Evaluation Method 36
3.5 Genetic Algorithm Parameter Settings 37
4 Results 39
4.1 Experimental Results and Analysis 39
4.2 Analysis of Optimized Weights for GA-based Rerankers 42
4.3 Limitations and Future Works 44
5 Conclusion 46
References 4
Comparative analysis of performance decline and anode degradation in polymer electrolyte membrane fuel cells under fuel starvation: A strategy for electrode condition assessment using full electrode equivalent circuit model and anode-separated distribution of relaxation times
No full textThis work investigates the impact of anode catalyst layer (CL) degradation in Polymer Electrolyte Membrane Fuel Cells (PEMFCs) under fuel starvation using accelerated stress tests (ASTs). Anode cycling ASTs simulate intermittent fuel starvation and reveal that reduced hydrogen oxidation reaction (HOR) kinetics drive predominant anode degradation. Cyclic voltammetry results show a significantly greater decline in electrochemical surface area in the anode CL compared to the cathode, emphasizing the impact of anode degradation on PEMFC performance decay. A rapid decrease in ionic resistance in low-humidity non-faradaic EIS and SEM images of aged MEAs reveals that structural changes in the anode carbon support structure are the primary degradation mechanism after ASTs. The increase in anode CL charge transfer resistance as the AST progresses aligns with anode impedance data fitting results using the Transmission Line Model and cathode impedance data fitting results using the Full Electrode Circuit Model. The distribution of relaxation times analysis on the separated anode circuit from cathode ORR impedance data indicates shifts in relaxation time, representing anode resistance changes. This work advances the understanding of anode degradation mechanisms under fuel starvation and provides insights into developing improved AST protocols and impedance analysis techniques to enhance PEMFC durability and performance. © 2025 Elsevier LtdFALSEsciescopu
LATP-incorporated Cellulose Membrane for Direct Lithium Extraction from Salt Lake Brine under an Electrochemical System
No full textLithium (Li) is a critical raw material for lithium-ion batteries (LIBs) and energy storage systems (ESS). With the ongoing clean energy transition, the demand for lithium is expected to rise continuously. However, conventional lithium extraction methods from salt lake brines, which rely on solar evaporation, face significant limitations in meeting this increasing demand and present environmental issues. Solar evaporation based lithium extraction typically requires 10–24 months and consumes substantial amounts of water, particularly in dry regions like the lithium triangle of South America. Due to these limitations, Direct Lithium Extraction (DLE), a method capable of selectively extracting lithium with high efficiency in a short time, has recently emerged as a promising alternative to evaporation-based methods. In this study, we developed an electrically driven lithium-selective membrane as a DLE technology due to its low energy consumption, high extraction efficiency, and environmental advantages, including its compatibility with renewable energy sources. The lithium-selective membrane, referred to as the LATP/CA membrane, incorporates the NASICON-type lithium ionic conductor LATP (Li1.3Al0.3Ti1.7(PO4)3) for high lithium selectivity. Additionally, cellulose acetate (CA) was used to maintain the mechanical strength of the membrane. Lithium separation tests were conducted under both binary systems and simulated brine solutions, along with a multi-step process, to examine the effects of applied current and multi-ion systems on membrane performance. Under an applied current of 200 µA in a simulated brine solution, the LATP/CA membrane achieved lithium selectivity ratios of Li/Mg: 467.3, Li/Na: 44.8, and Li/K: 22. This lithium selectivity could be attributed to the crystal structural characteristics of LATP. Consequently, the application of the LATP/CA membrane shows the potential to address supply challenges associated with conventional evaporation methods while mitigating environmental concerns for lithium extraction.MasterAbstract ⅰ
Contents ⅱ
List of Figures ⅳ
List of Tables ⅴ
Ⅰ. Introduction 1
Ⅱ. Experimental sections 5
2.1. Materials 5
2.2. Membrane fabrication 5
2.3. Characterization 5
2.3.1. Synchrotron high-resolution powder diffractometer 5
2.3.2. Attenuated total reflectance Fourier transform infrared spectroscopy 5
2.3.3. Ultra-high resolution field emission scanning electron microscope 6
2.4. Lithium separation test 6
2.5. Stability test for LATP/CA membrane 6
2.6. Calculations 7
2.6.1. Calculation of lithium separation performance 7
2.6.2. Calculation of weight change 7
Ⅲ. Results and discussions 9
3.1. Membrane fabrication and characterization 9
3.2. Lithium separation performance for binary solutions 13
3.3. Lithium separation performance for simulated brine conditions 15
3.4. Two-step lithium separation process 19
3.5. Stability of LATP/CA membrane 24
Ⅳ. Conclusions 25
Ⅴ. References 2
A phase-optimized NiFe-LDH/NiB heterostructure as an efficient and durable oxygen evolution electrocatalyst in alkaline media
No full textNickel-iron layered double hydroxide (NiFe-LDH) has attracted considerable attention as an efficient electrocatalyst for the oxygen evolution reaction (OER) in alkaline media. However, the irreversible phase transition from gamma-Ni(Fe)OOH to beta-Ni(Fe)OOH, which is based on the low thermodynamic stability of gamma-Ni(Fe)OOH, results in the poor durability of NiFe-LDH. To address this, this study designs an NiFe-LDH/NiB heterostructure (NiFe@NiB). Because NiB acts as an electron acceptor, it modulates the Ni oxidation state (Ni3+ -> Ni(3+delta)+) and facilitates the beta-to-gamma phase optimization. Notably, NiFe@NiB maintains a higher gamma-phase fraction during OER cycling and exhibits an expanded 2D layered structure, which is a structural feature of the active gamma-phase. In conclusion, NiFe@NiB requires 75 mV lower overpotential to achieve 10 mA cm-2 and one-fifth degradation rate with 93.2% reduced Fe leaching over 120 hours of durability test compared to NiFe-LDH. This work presents a compelling strategy for designing efficient and durable electrocatalysts for sustainable hydrogen production.FALSEsciescopu
DESIGN AND IMPLEMENTATION OF VIRTUAL FITTING SYSTEM ON THE WEB
No full textSince the coronavirus pandemic, online shopping has become a global phenomenon, and this purchasing behavior is becoming common. A virtual tryon system allows customers to virtually try on clothes, accessories, makeup or other products before purchasing them. Previous research generally focuses on maintaining the characteristics (e.g., texture, logo, and embroidery) of clothing images when warping them into arbitrary human poses. This study investigated and analyzed existing research on the virtual tryon model and evaluated the applicability of the virtual tryon model under various conditions in a realistic online shopping environment. We adopted the Adaptive Content Generating and Preserving Network (ACGPN) model. The experimental conditions were evaluated considering the background of the customer’s photo, the type of clothing already worn, the type of clothing selected, and the customer’s posture. The results of the experiment showed that the outcomes of warping the clothing were different depending on the conditions, and that they were reacting sensitively. © 2025, ICIC International. All rights reserved.FALSEscopu