34254 research outputs found
Sort by
Synthesis of Fluoroether-Based Electrolytes for High-Voltage Lithium Batteries
No full textThis thesis synthesized two types of fluorinated ether electrolytes using six different fluorinated alcohol compounds as starting materials. Sodium hydride was used to remove the hydrogen from the fluorinated alcohol compounds, forming negatively charged oxygen. Subsequently, the generated oxygen with a negative charge reacted with alkoxy halides or methyl iodide through SN2 reaction. This resulted in the formation of fluorinated ether compounds with varying degrees of fluorination and ether chain lengths, allowing for the adjustment of the electrolyte's ionic conductivity and oxidative stability to achieve a better balance. We also optimized the stoichiometry and solvent conditions. The synthesized compounds were then used as additives (0.2%) in Hopax electrolytes for electrical testing. Different rate charge-discharge tests were conducted on lithium nickel manganese oxide cathode
Synthesis of Two-Dimensional Porous Polyimide-Covalent Organic Framework Materials
No full textThe content of this thesis mainly focuses on the synthesis of two-dimensional porous polyimide-covalent organic framework (PI-COF) materials. By using a known dianhydride, pyromellitic dianhydride (PMDA), as a starting material, two different amine monomers, 24 and 27, with distinct central structures are employed to construct two types of two-dimensional porous covalent organic framework materials, 28 and 29, with different molecular structures. The obtained porous polyimide-covalent organic framework materials were then subjected to tests such as specific surface area analysis and thermogravimetric analysis to investigate the structures and properties
Study of Platinum Decorated Indium Oxide Nanoflowers for Acetone Gas Sensing Applications
No full textThis study focuses on the sensing properties of indium oxide (In2O3) and platinum (Pt) decorating samples towards acetone gas. The growth of In2O3 is carried out using a hydrothermal method. Different nanostructure morphologies are observed by varying the quantity of sodium dodecyl sulfate (SDS) and the growth time. Pt with various appearances decorated on In2O3 is achieved by adjusting the dihydrogen hexachloroplatinate (IV) hexahydrate concentration. The physical and electrical properties are measured to investigate the correlation between morphological changes and the sensing performance of the sensor toward acetone gas.
This research utilizes noble metal, which can enhance the material's conductivity and act as a catalyst at low temperatures, to improve sensing performances. According to the experiment results, In2O3 nanoflowers synthesized with 0.192 g SDS for 10 hours exhibit the highest responsivity. Furthermore, the sensing performance is further enhanced by Pt decoration at a concentration of 0.25 wt.%. The research indicates that undecorated In2O3 has a responsivity of 41.51% toward 100 ppm acetone gas under a working temperature of 250 \uc2\ub0C. After 0.25 wt.% Pt decoration, The In2O3 sample lowered its working temperature to 200\uc2\ub0C with a higher responsivity of 57.11%. The Pt/In2O3 acetone gas sensor demonstrates response and recovery times of 14.3 and 11.6 seconds, respectively. Meanwhile, it also has high selectivity to acetone gas
The study of sulfonated polymer blended with polyethylene glycol filled in expanded poly(tetrafluoroethylene) composite membrane for hydrogen fuel cells
No full textIn this study, the sulfonated polyarylether polymer (SP4) previously developed by the laboratory was dip-coated the commercially available hydrophilic modified backing material ePTFE. After adding the backing material, it was found that the proton conductivity of the composite membrane would decrease 20~25%, so this study uses polyethylene glycol (PEG) as a cross-linking agent blended with the sulfonated polyaryl ether polymer to improve the proton conductivity of the membrane.
The blended with 0.1% and 0.5% is found to be the best ratio after measuring the proton conductivity of the blended with PEG series membranes, and its proton conductivity can achieve 228.31~235.63 mS/cm under the environment of 80oC and 95%RH. The pure membrane SP4 can effectively increase (187.37 mS/cm), so the ePTFE series membranes will be blended with 0.1% PEG and 0.5% PEG to improve the properties of the composite membrane.
The structural changes of the ePTFE series composite membranes after filling were observed by SEM, and the results showed that all sulfonated polymers were dip-coated inside the membranes. At 80 oC, the dimensional change rate of the composite membranes is < 16%, which can effectively reduce the dimensional change of the film at high temperature compared with the pure membrane SP4 (28.7%). In terms of proton conductivity, in the environment of 80 oC and 95%RH, the proton conductivity of SP4+0.5%PEG+ePTFE is 192.07 mS/cm, which is better than pure membrane SP4 (187.37 mS/cm), Nafion 211 ( 113.75 mS/cm). In terms of Single Cell Efficiency Measurement, the efficiency of SP4+0.5%PEG+ePTFE is 0.88 W/cm2, which is close to that of pure membrane SP4, which is 0.9 W/cm2. Based on the above, adding cross-linking agent PEG can effectively improve proton conductivity, and adding support material It can effectively improve the dimensional change rate at high temperature, so the modified composite membrane can have an efficiency close to that of the pure membrane in the single cell test
Design and Implementation of SRAM-Based Anti-Noise PUF Device Authentication Circuit
No full textIn today's era, smart electronic products that we use every day are closely related to chips. Daily life is also closely linked to various smart devices with centralized management, such as the Internet of Things, smart homes, health monitoring, self-driving cars, and even data monitoring in high-pressure environments, all of which involve many edge devices, which also create potential security risks. However, edge devices in the Internet of Things are usually limited by budget and performance, so they are not suitable for using complex and expensive encryption schemes. Using PUF (Physical Unclonable Function) not only has a relatively low cost but also avoids the old method of storing keys in Non-Volatile Memory (NVM), which is easily invaded and cracked to obtain the key.
PUF is a circuit that generates electronic fingerprints based on slight differences in the semiconductor process. Its operation principle is based on the physical properties of hardware, such as the capacitance and resistance of the circuit in the chip. These properties will have slight and non-replicable differences in different devices, thus achieving good uniqueness. The SRAM PUF used in this paper is designed based on the memory commonly configured in various edge devices, where each SRAM storage unit is considered as a random tree generator, and the generated random bit string is called a PUF response. Due to the differences in the manufacturing process of each SRAM unit, the complexity of the PUF response is unique and cannot be replicated among different SRAM PUF chips.
However, SRAM PUF faces many challenges, such as storage aging, leading to unstable PUF responses. Therefore, to achieve a reliable SRAM PUF design, additional auxiliary circuits need to be designed to stabilize and compensate the data.
This paper combines anti-aging and compensation mechanisms to achieve a reliable device authentication scheme. The design has achieved good results in reducing PUF response distortion and increasing uniqueness through experimental and simulation verification. This design can be applied to device authentication and encryption applications for edge devices, with the advantages of low cost, improved security, and stability
Generalized Block-Based Spatial Modulation based Reconfigurable Intelligent Surface assisted MIMO with Low Complexity Detector
No full textMultiple input multiple output (MIMO) technology serves as an alternative method to improve the spectral efficiency (SE) for the system. Nonetheless, a higher count of transmit antennas results in elevated power consumption and greater system complexity. As a result, spatial modulation (SM) has emerged as a popular technique in MIMO systems, providing the advantages of reduced power consumption and increased transmission capacity through the spatial domain. Generalized block based spatial modulation (GBSM) is a MIMO technology, in which the antennas activated by each time slot will be represented in the block. The presence of tall buildings and other structures in urban areas poses a significant challenge in wireless communication, often resulting in signal transmission blockage between the user equipment (UE) and base station (BS). To address this challenge effectively, reconfigurable intelligent surface (RIS) have recently emerged as a remedy, propagational control techniques can address coverage and signal quality limitations. Combining RIS and GBSM technologies in a joint framework offers potential for high energy and spectral efficiency. In this thesis, we propose a MIMO model based on RIS-enabled GBSM
Development of a Low Insertion Loss Biosensor Array for Detection of Lung Cancer Using MEMS Technology
No full textTaiwan's Ministry of Health and Welfare released the National Cancer Ranking Report in 2021, there are about 13,000 lung cancer cases in my country every year, and lung cancer is also recognized as one of the cancers with the highest mortality rate in the world. Lung cancer can be classified into two major types based on differentiation or location: Small Cell Lung Cancer (SCLC) and Non-Small Cell Lung Cancer (NSCLC). In clinical practice, the biomarker for detecting SCLC is neuron-specific enolase (NSE), while the biomarker for detecting NSCLS is cytokeratin fragment 19 (CYFRA21-1). Higher concentrations of these two tumor markers in the blood indicate a higher likelihood of having lung cancer. Early detection enables individuals to promptly seek medical examination and treatment at the hospital. This Thesis focuses on the development of a miniaturized lung cancer array chip capable of simultaneously detecting CYFRA21-1 and NSE. The goal is to apply this chip in future relevant rapid diagnostic devices.
The developed Flexural-plate Wave arrayed sensing device in this research paper involves five steps of thin film deposition and five steps of lithography process. Utilizing Microelectromechanical Systems (MEMS) technology, an FPW array chip is developed, and the Self-Assembled Monolayers (SAMs) technique is employed to immobilize CYFRA21-1/NSE antibodies on the sensing region at the back of the device. This completes the development of an arrayed microsensor for lung cancer capable of simultaneously detecting CYFRA21-1 and NSE antigen concentrations. Furthermore, the thesis explores the optimal logarithmic value for the Reflective Grating Structure (RGS) to reduce insertion loss and improve the sensitivity of mass sensing. The study also focuses on important sensing characteristics such as linearity and specificity of the device.
According to the experimental results, the developed FPW device in this thesis exhibits a minimum insertion loss (-27.524 dB). The mass sensitivity is measured to be 121.954 cm2/g, and the sensing linearity (R2) is as high as 0.994. The FPW biosensing chip developed in this research paper shows a linear measurement range of 0.5 to 4 ng/ml for CYFRA21-1 and 1 to 25 ng/ml for NSE. These ranges cover the normal standard values in the human body (2.1 ng/ml for CYFRA21-1 and 12.5 ng/ml for NSE). The measurement linearity (R2) is found to be 0.988 and 0.987 for CYFRA21-1 and NSE, respectively. The detection limits are determined to be 0.443 and 0.743, respectively. Regarding the specificity measurement, this thesis demonstrates that the frequency shifts of two other tumor markers are less than 7% of the frequency shift of CYFRA21-1/NSE, indicating high specificity.
In conclusion, the arrayed lung cancer sensing chip developed in this thesis demonstrates excellent sensing characteristics, which will contribute to the future development of related microsensing systems
A Study of M&A Classifier based on Patent and Finance Metrics
No full textThis study utilizes machine learning techniques to predict M&A transactions in the fintech industry, adopting a financial and technical model approach to anticipate potential M&A targets and assess post-merger performance and innovation. Besides integrating financial and technical indicators, the study also introduces variables such as cultural fit, potential fit, and similarity fit to enhance predictive capabilities and delve deeper into factors affecting M&A outcomes. The findings reveal the knowledge-based model's superiority over the financial-based one, highlighting the significance of patent-related indicators and similarity variables as key in model training. Moreover, the number of events a company participates in significantly impacts post-merger performance. Ultimately, this study offers valuable decision-making tools for senior management, incorporating financial and patent data to better understand key factors in M&A, providing directions and suggestions for further research, and facilitating a more in-depth understanding of the M&A field both academically and practically
Research on GaxZn1-xO/ZnO Heterojunction Thin-Film Transistors with Negative Differential Resistance Phenomenon
No full textThis paper investigates the use Mist-CVD to fabricate HTFTs and measure their electrical properties, focusing on observing the Negative Differential Resistance (NDR) . In this study, the main mechanisms responsible for the NDR are the Real-space transfer and the Gunn effect. HTFTs are constructed with a ZnO thin film as the active layer and a Ga0.10Zn0.90O thin film as the barrier layer to facilitate electron transfer when a gate bias is applied. Subsequently, NDR is observed in the devices during the measurement of ID-VD characteristics at 25\uc2\ub0C and 200\uc2\ub0C.
All HTFTs employ ITO thin films as the gate electrode. Aluminum metal is used as the source and drain electrodes, and the active layer consists of ZnO. Under the mentioned conditions, the barrier layer is composed of GaxZn1-xO thin film, and an additional layer of Ga2O3 thin film is inserted between the active layer and the barrier layer as a spacer layer. The role of the spacer layer is to enhance the potential well effect in the barrier layer when gate bias is applied, making it less likely for electrons that have already undergone real-space transfer to move back into the active layer.
By varying gallium content in the GaxZn1-xO barrier layer (5%, 10%, 15%) and observing electrical measurements, it is found that 10% gallium content exhibits the optimal peak to valley current ratio (PVCR) at 200\uc2\ub0C. Based on the best PVCR performance of Ga0.10Zn0.90O as the barrier layer, the thicknesses of the active layer and the barrier layer are adjusted to observe NDR. Finally, the study reveals that at 25\uc2\ub0C, the device structure Ga0.10Zn0.90O (20 nm)/ZnO (20 nm) achieves the best PVCR of 2.13. At 200\uc2\ub0C, Ga0.10Zn0.90O (20 nm)/Ga2O3 (5 nm)/ZnO (20 nm) shows the best NDR, with a PVCR of 3.3. This thesis aims to investigate the mechanisms and electron flow trends when HTFTs exhibit NDR
CLVD: Cost-aware load-balanced VNF deployment in data center networks
No full textNetwork function virtualization (NFV) is an emerging network technology. Different from the traditional network architecture, where network services must be installed on dedicated hardware devices, NFV implements these network functions through software. NFV deploys them on the infrastructure to become virtual network functions (VNFs). The function of linking these VNFs together to achieve specific network functions is called a service function chain. Through the NFV technology, network providers can provide network services more efficiently and flexibly and reduce the cost of hardware equipment. How to deploy VNFs on physical machines is a problem deserving investigation. Since VNF deployment will greatly affect system performance, it is important to design an effective deployment method. In past studies, some methods aimed to maximize the number of deployed VNFs and minimize the number of physical hosts required. However, they may cause load unbalanced between physical machines. On the other hand, though some studies considered load balance, but they resulted in high communication costs for VNFs.
In view of this, the thesis proposes a cost-aware load-balanced VNF deployment (CLVD) scheme, whose objective is to balance the loads between physical machines while minimizing the communication cost. We calculate a deployment expected value that considers the residual resource, communication cost, workload, and the load of neighboring physical machines. This expected value is used as a reference to select a suitable physical machine to deploy each VNF. Moreover, to avoid overloading some physical machines, CLVD also performs VNF migration to share the load of a physical machine with others. Through simulations, we show that our proposed CLVD scheme can balance the loads of physical machines and reduce the communication cost, as compared with other methods