Browsing by Author "Tsai, Cheng-Ming"

Now showing 1 - 6 of 6

An Energy-Band Model for Dual-Gate-Voltage Sweeping in Hydrogenated Amorphous Silicon Thin-Film Transistors
(2019-06-01)
Chen, Guan-Fu
;
Chen, Hong-Chih
;
Chang, Ting-Chang
;
Huang, Shin-Ping
;
Chen, Hua-Mao
This paper clarifies the correct transmission mechanism, misattributed in the previous research, of a hydrogenated amorphous silicon thin-film transistor (a-Si:H TFT) device. Complete drain current-gate voltage (I-D-V-G) transfer characteristics including the forward and backward gate sweeps (gate voltage carried out in the OFF-state -> on-state and the ON-state -> OFF-state) are performed, and the physics models of an energy-band schematic are clearly explained. This paper reveals that the I-D-V-G curve stretch-out behavior of the subthreshold region is due to the leakage of the Poole-Frenkel region. At the Poole-Frenkel regions, electrons transfer to the drain terminal, and holes remain in the valance band of the a-Si:H bulk. The remaining holes represent the positive voltage and cause the electron barrier height lowering. This causes an increase in subthreshold current of the subthreshold region. Finally, three experiments are performed to prove the correctness of these models, and the a-Si TFTs device improvement methods will be proposed to enhance the stability and performance of product of the a-Si TFTs.
Drain-Induced-Barrier-Lowing-Like Effect Induced by Oxygen-Vacancy in Scaling-Down via-Contact Type Amorphous InGaZnO Thin-Film Transistors
(2018-01-01)
Yang, Chung-I.
;
Chang, Ting-Chang
;
Liao, Po-Yung
;
Chen, Li-Hui
;
Chen, Bo-Wei
This investigation considers a method to ameliorate drain induced barrier lowing behavior in amorphous-indium-gallium-zinc-oxide thin-film transistors. The Vth is found to shift negatively when increasing the ID-VG measurement condition VD from 0.1 to 15 V. The current-voltage curves show that this degradation is caused by the effective channel length (L-eff) being shorter than the mask channel length (L). Using the transmission line method to extract Leff, we discover that the degradation will be completely suppressed by an annealing treatment. As a result, the degradation mechanism of shorter channel length a-IGZO thin film transistors is due to oxygen-vacancies which are located between the channel and the source/drain junction.
Enhanced Light Harvesting with a Reflective Luminescent Down-Shifting Layer for Dye-Sensitized Solar Cells
(2013-06-26)
Hosseini, Zahra
;
Huang, Wei-Kai
;
Tsai, Cheng-Ming
;
Chen, Teng-Ming
;
Taghavinia, Nima
For a dye-sensitized solar cell with a near-infrared squaraine (SQ1) sensitizer, the photovoltaic performance was enhanced remarkably with a reflective luminescent down-shifting (R-LDS) layer to increase the light-harvesting efficiency at the wavelength region 400-550 nm where the SQ1 dye has weak absorption. Relative enhancements greater than 200% in IPCE near 500 nm and 40-54% in J(SC) were achieved with red phosphor CaAlSiN3:Eu2+ as the LDS material, attaining 5.0 and 4.8% overall efficiencies of power conversion for the R-LDS layer coated on the counter electrode (front illumination) and working electrode (back illumination), respectively.
Hybrid Titania Photoanodes with a Nanostructured Multi-Layer Configuration for Highly Efficient Dye-Sensitized Solar Cells
(2013-05-02)
Wu, Hui-Ping
;
Lan, Chi-Ming
;
Hu, Jyun-Yu
;
Huang, Wei-Kai
;
Shiu, Jia-Wei
To construct a hybrid titania photoanode containing nanoparticles and nanorods of varied size in a multilayer (ML) configuration for dye-sensitized solar cells, the essence of our ML design is a bilayer system with additional layers of nanorods of well-controlled size inserted between the transparent and the scattering layers to enhance the light harvesting capability for photosensitizers with small absorptivity, such as Z907. We measured charge-extraction and intensity-modulated photoelectric spectra to show the advantages of one-dimensional nanorods with an improved electron-transport property and an upward shift of the potential band edge; a favorable ML configuration was constructed to have a cascade potential feature for feasible electron transport from long nanorods, to normal nanorods, to small nanoparticles. On the basis of the ML system reported herein, we demonstrate how the performance of a Z907 device is improved to attain eta similar to 10%, which is a milestone for its future commercialization.
Investigating degradation behaviors induced by hot carriers in the etch stop layer in amorphous InGaZnO thin film transistors with different electrode materials and structures
(2017-12-31)
Yang, Chung-I
;
Chang, Ting-Chang
;
Chen, Bo-Wei
;
Chou, Wu-Ching
;
Liao, Po-Yung
This work investigates the hot carrier effect in via-contact type amorphous indium gallium zinc oxide thin film transistors with various source/drain materials and structures. According to previous research, the redundant drain electrode plays an important role in hot carrier stress-induced degradation, which leads to carrier-trapping in the etch stop layer between the active layer and the redundant drain electrode. Hot carrier stress has different influences on device characteristics, depending on materials and structure. Hot carrier stress causes more electron trapping in the etch stop layer below the redundant drain electrode in the presence of smaller source/drain metal work function or a longer redundant drain electrode. To further verify the mechanisms of the degradation behavior, the barrier height for Fowler-Nordheim-tunneling is extracted by a fitting charge trapping model. It is found that the barrier height for Fowler-Nordheim-tunneling is different for different source/drain materials.
On Accelerating Multi-Layered Heterogeneous Network Embedding Learning
(2018-01-01)
Shuai, Hong-Han
;
Tsai, Cheng-Ming
;
Hsu, Yun-Jui
;
Hsiao, Ta-Che
;
電機工程學系
With the rapid growth of online social networks and IoT networks, mining valuable knowledge from the graph data become important. Meanwhile, as machine learning algorithms show their powers in prediction, different machine learning algorithms are proposed for different applications, e.g., personal recommendation, price prediction, communication anomaly detection. However, it is challenging to extract network features from graph data as the inputs for machine learning algorithms. One of the promising approaches is to use graph embedding approach, which extracts the valuable information of networks from each node into low dimensional vectors. However, the graph embedding approaches on a large-scale network require tremendous training time. Therefore, in this paper, we propose NOde Differentiation for Graph Embedding (NODGE) to prioritize the nodes, while high priority nodes are allocated with more resources to train their representations. We also theoretically analyze the proposed NODGE. Experimental results show that the proposed method reduces the training time of state-of-the-art method by at least 30.7%.